Introduction

Hospitalization is a time of significant change for patients with type 2 diabetes (T2DM). Although 72% of patients take oral medication(s) to manage their T2DM as outpatients [1], current guidelines and consensus statements recommend holding noninsulin therapy at the time of hospital admission and using insulin for achieving glycemic targets during hospitalization [2,3,4,5]. Patients with T2DM who are using insulin therapy prior to admission typically require adjustment of their insulin regimen, either with insulin type or with dose (or both). This is done to avoid hyperglycemia and hypoglycemia in the acute care setting, and oftentimes because the patient’s home insulin type is not available on the inpatient formulary. Patients newly diagnosed with T2DM require initiation of glucose lowering therapy as well as diabetes self-management education.

The time of transition from the inpatient to outpatient setting for patients with diabetes, regardless of the primary reason for hospitalization, is frequently associated with preventable adverse events and hospital readmission [6, 7]. Conditions that caused alteration in glycemic control during hospitalization are often not fully resolved by the time of hospital discharge. Both inpatient and previous outpatient diabetes regimens may be too aggressive or not aggressive enough as patients continue to recover from their acute illness or surgery at home and return to their usual diet and activity level [7]. In a survey of 19 US hospitals, Rodriquez and colleagues found that only a quarter of hospitals with established glycemic management programs had a standardized protocol for providing patients with diabetes education and follow-up instructions regarding medications and continuity of care [8]. This number is likely lower when all hospitals are considered since the hospitals surveyed had well-established glycemic control teams.

In 2012, >25% of all inpatient hospitalization days were incurred by people with diabetes [9]. Patients with diabetes, regardless of the primary reason for hospitalization, are at increased risk for readmission compared to those without diabetes [10•, 11]. Using a patient-centered approach to optimize therapy, anticipate needs, provide education, and ensure continuation of care following hospitalization has the potential to reduce hospital readmissions for many patients with T2DM. The objective of this review is to provide clinicians with practical evidence-based recommendations for transitioning hospitalized patients with T2DM to home.

Optimizing the Patient’s Discharge Diabetes Regimen

Hospitalization can serve as an opportunity to identify and initiate treatment for patients with previously undiagnosed T2DM or optimize the treatment regimen of patients with poorly managed T2DM [4]. Unfortunately, several studies have shown that outpatient diabetes medications are frequently discontinued during hospitalization and not restarted at time of discharge [12, 13]. In a study of patients with T2DM who were treated with insulin prior to and during hospitalization, those who continued insulin therapy at discharge had lower rates of hospital readmission, similar rates of hypoglycemia (defined as an ICD-9 diagnosis code), and improved survival during the 12 months following hospitalization [14]. Despite evidence demonstrating that intensification of the diabetes regimen at hospital discharge in patients with elevated A1c is associated with fewer readmissions [15, 16], adding new diabetes medication or altering the prior outpatient regimen occurs infrequently [17,18,19,20].

Only one prospective study has evaluated strategies for designing a discharge diabetes regimen for patients with T2DM. Umpierrez and colleagues designed and evaluated an algorithm for selecting discharge diabetes medications based on admission A1c for T2DM patients taking no diabetes medications, oral antidiabetes (OAD) agents, or ≤0.4 units/kg insulin prior to hospitalization [21••]. In this study, patients admitted with an A1c <7% were discharged on their pre-admission medications (oral agents and/or insulin), those with an A1c between 7 and 9% were discharged on their pre-admission oral agents plus 50% of the inpatient basal insulin dose, and patients with an A1c >9% were discharged on the combination of their pre-admission oral agents plus 80% of the inpatient basal insulin dose or on a basal-bolus insulin regimen. After 12 weeks, the A1c was reduced by 0.1% in patients with a baseline A1c <7%, 0.8% when baseline A1c 7–9%, and 3.2% when baseline A1c >9%. The percentage of patients who experienced hypoglycemia defined as a blood glucose (BG) <70 mg/dL varied according to therapy in the patients who were discharged on OAD (22%), basal insulin alone (25%), OAD plus basal insulin (30%), and basal-bolus insulin (44%). Although few patients experienced severe hypoglycemia (defined in this study as BG <40 mg/dL), the risk of hypoglycemia that is associated with intensifying the patient’s diabetes medications by adding insulin at discharge is potentially worrisome. Of note, the American Diabetes Association (ADA) now defines “clinically significant hypoglycemia” as a BG <54 mg/dL and the glucose alert value is defined as BG ≤70 mg/dL [2].

The goal of any diabetes regimen is to effectively achieve and maintain glycemic targets while minimizing side effects such as hypoglycemia [5, 22•, 23•]. Regardless of severity, symptomatic hypoglycemia is associated with a higher risk for cardiovascular events, hospitalization, and mortality in patients with T2DM [24,25,26,27]. In addition, patients who experience hypoglycemia report higher rates of depression and lower quality of life [28, 29]. In a telephone survey of patients discharged on long-acting insulin conducted 1 week following hospital discharge from an academic medical center in California, 30% of patients reported at least one episode of BG <70 mg/dL [30]. Wei and colleagues found that hypoglycemia occurred in 46% of patients with an average A1c of 10.3% discharged from the hospital with insulin therapy [31•]. Hypoglycemia related to insulin and other diabetes medications is a leading cause of emergency department visits for adverse drug events among older adults [32].

Additional randomized controlled studies are needed to determine how to effectively intensify diabetes regimens after hospitalization while minimizing the risk of hypoglycemia [33, 34]. Table 1 provides a suggested approach for optimizing diabetes medications for patients with T2DM at hospital discharge based on both the A1c [21••] and the most recent American Diabetes Association/European Association for the Study of Diabetes (ADA/EASD) position statement and the American Association of Clinical Endocrinologist (AACE) consensus statement for pharmacological management of patients with T2DM [22•, 23•]. Patients who have poorly controlled diabetes at the time of hospital admission require intensification of their outpatient regimen at hospital discharge, but this intensification does not necessarily mean that all patients require continuation of the insulin regimen started during hospitalization [16, 20]. The suggestions for insulin dosing (50–80% of the inpatient regimen) included within Table 1 are based on the algorithm developed by Umpierrez and colleagues [21••] and assumes that insulin has already been titrated in the inpatient setting to achieve good glycemic control given the patient’s current clinical status.

Table 1 Suggested approach to prescribing diabetes medications for patients with T2DM at hospital discharge
Full size table

Use of correction-dose insulin to supplement either basal only or basal-bolus regimens after hospital discharge is not generally recommended, especially at bedtime. In the hospital setting, correction-dose insulin may be helpful when used in combination with basal insulin to correct hyperglycemia in the patients who are not eating or as part of a temporary dose-finding strategy in insulin-naïve patients during times of unknown or rapidly changing insulin requirements [4, 35]. Correction-dose insulin in combination with basal insulin is also an effective alternative to basal-bolus insulin regimens for the initial management of hyperglycemia in hospitalized patients with T2DM who were previously well-managed with noninsulin therapies or low doses (<0.4 units/kg/day) of insulin [36]. However, there is no evidence to support the safety or efficacy of correction-dose insulin in the outpatient setting. Patients with T2DM who require insulin at time of hospital discharge should be managed with scheduled doses of insulin that are then adjusted in the outpatient setting based on blood glucose patterns. Patients who are proficient at carbohydrate counting can be discharged to home on mealtime insulin doses that are based on an established insulin-to-carbohydrate ratio; however, initiating advanced carbohydrate counting is usually not feasible in the inpatient setting.

The selection of pharmacotherapy for glycemic management in the outpatient setting should take patient factors into consideration such as ability to follow recommendations, availability of support systems, and cognitive status as a way of ensuring patient safety [37]. Patients who are struggling to master safe and accurate insulin administration and injection technique may benefit from consideration of noninsulin options when discharge is imminent. Unless the patient has new symptomatic hyperglycemia [38], is at risk for ketoacidosis (DKA), or has underlying conditions (i.e., infections) that indicate a need for ongoing insulin therapy, noninsulin options may be preferred.

The decision to initiate insulin or to add mealtime insulin to existing basal insulin at discharge should be determined well in advance of the day of discharge to ensure that there is time to assess patients’ ability to afford insulin as a treatment regimen and their ability to perform the necessary insulin administration and glucose monitoring skills [39, 40]. For some patients with physical or cognitive issues that interfere with safe insulin practice, the risks associated with insulin regimens may outweigh potential benefits.

Patients with visual impairment or issues with fine hand coordination often find insulin pens easier to use than vials and syringes because they do not need to draw up the insulin and remove air bubbles from the syringe. However, insulin pens require the ability to attach a pen needle before each injection, and remove the pen needle after the injection. Patients with tremor, cognitive impairment, or poor eyesight often have difficulty with these steps. Use of safety pen needles may be a solution for patients with dexterity issues; however, the priming step (air shot) that is required before dialing the dose with either a regular or a safety pen needle can still be challenging. Unless the patient has the opportunity to practice insulin self-administration in the hospital prior to discharge, there is no way to identify these potential issues and offer accommodations. For example, in situations where a patient is unable to independently put a needle on an insulin pen or draw insulin up in a syringe, it may be possible to train a family member or friend to assist the patient.

Anticipating Patients’ Needs During the Immediate Post-discharge Period

Discharge planning should begin at the time of hospital admission [2, 4]. The patient’s cognitive ability, current knowledge, health literacy level, numeracy skills, family support, and financial resources should be assessed for all patients with T2DM [4, 8]. As noted above, for patients who will be discharged to home and who are new to insulin, visual acuity and dexterity also need to be assessed. These assessments should take place on the day of hospital admission or as soon as the patient’s condition is stable [41]. Early assessment allows sufficient time to address problem areas and secure additional resources and referrals that can be made available at time of discharge.

Appropriate discharge planning must also include assessment of potential barriers to following the prescribed therapy and obtaining post-discharge ongoing care. Nearly 50% of patients discharged from an urban hospital reported financial restraints, lack of health insurance, and lack of transportation that affected their ability to obtain follow-up care [42]. Rubin and colleagues conducted semi-structured interviews of 20 patients with diabetes who were readmitted within 30 days of a previous hospitalization to explore factors that had contributed to the readmission [43•]. They found that more than 50% of patients reported difficulty getting to follow-up medical appointments. Others reported not having sufficient help with nursing care, medications, and meals; 10% of patients were unable to obtain prescriptions for insulin following discharge. Readmissions among patients with diabetes are often related to social and economic issues such as patients’ inability to obtain glucose monitoring supplies or diabetes medications [30, 44].

Involving the patient and patient’s family and ensuring that they are comfortable with and able to follow discharge plans needs to be a priority. All medication changes need to be clearly communicated to the patient and included in the discharge summary that is shared with the patient’s outpatient providers. Armor and colleagues found that 81% of patients who had their medications reconciled by a pharmacist in a family medicine clinic within 10 days of hospital discharge had at least one medication discrepancy between the patient’s current medications and the list provided at hospital discharge [45]. As part of the discharge process, patients should be asked to repeat the generic and brand names, doses, and administration times for all medications and describe which medications taken prior to admission should be stopped or changed. Patients should learn the generic and brand name of their medications to reduce errors (e.g., taking metformin AND Glucophage® at the same time). Instructions should not include the words “continue all home medications” since this can lead to confusion, especially if medication reconciliation on admission was not adequately performed.

If a patient will be discharged on insulin, other diabetes medication(s), and/or glucose monitoring supplies that they have not previously used at home, an assessment of the patient’s ability to afford these prescriptions should be performed before the patient leaves the hospital. Even patients who have health insurance may be underinsured, have high out-of-pocket costs for prescriptions, or difficulty obtaining prescriptions due to lack of formulary coverage for the specific brand or type (vial vs. pen) of insulin for which the prescription is written. Some hospitals provide patients with a 30-day supply of their outpatient medications at discharge for a reduced or no cost [46]. At other institutions, care coordinators (available to address these kinds of issues, for example, social workers, nurses, pharmacists, or diabetes educators) work with retail pharmacies to provide patients with information regarding the cost of their prescriptions before they are actually filled. If the cost is prohibitive, this allows time for the inpatient team to find less expensive therapeutic options before the patient leaves the hospital.

Providing patients with clearly written discharge instructions, in a language they are comfortable reading, that include all the components listed in Fig. 1 helps reinforce verbal instructions [47]. These discharge instructions should be delivered in a concise, simple, standardized, and easy to understand manner, and include a contact person or clinic that can be called if the patient has questions or concerns and does not already have an established diabetes provider. Table 2 is a checklist that can be completed throughout the hospitalization to ensure that patients know and have everything needed for a successful transition to home.

Fig. 1
figure 1

Example of standardized yet personalized written diabetes discharge instructions

Full size image
Table 2 Checklist for patients with type 2 diabetes who are being discharged to home.
Full size table

Providing Diabetes Education

Diabetes self-management education (DSME) improves health outcomes and glycemic control in patients with T2DM [48]. DSME is recommended at the time of new diagnosis of T2DM, annually, when new factors (such as diabetes complications, physical limitations, and social issues) influence self-management, and when care transitions such as hospital discharge occur [49]. Patients with pre-existing diabetes (diagnosis pre-dating hospital admission) who perceived that they did not receive effective self-management education prior to hospital discharge reported more difficulty at home and were more likely to have unplanned provider or emergency department visits or hospital readmissions within 30 days of hospital discharge [50•].

Hospitalization should serve as an opportunity to provide or reinforce DSME principles [51]. Unfortunately, teaching diabetes self-management skills in the inpatient setting is challenging because patients often do not feel well enough to engage in education, are under stress related to their hospitalization, and are in an environment with noise (e.g., alarms) and frequent interruptions that is not conducive to learning [29, 35]. Tests and procedures often take patients out of their rooms for a significant portion of the day, and patients may have inadequate rest, altered mental status, or poor concentration due to pain, pain medications, and other distractions [52•]. Even when patients are taught on the day of discharge, when they should feel better than any other time during the hospitalization, they often do not recall things they are taught [53]. In two randomized studies that assessed the impact of an inpatient diabetes education intervention, the investigators had low enrollment because many patients were too “overwhelmed by hospitalization” or in too much pain to participate [52•, 54]. Given today’s short hospital lengths of stay (LOS) and the inability to bill for inpatient education services, delivering comprehensive DSME in the inpatient setting is not practical in many hospitals [55].

To avoid overwhelming patients with too much information at a time when their capacity to learn and absorb new material may be limited, diabetes-related education in the hospital setting should focus on basic “survival skills” [4]. Although there are some differences across programs in the number of skills included [56, 57••, 58], Table 2 summarizes the basic survival skills that must be discussed prior to discharge [2].

Dungan and colleagues identified inpatient diabetes education that begins prior to the day of discharge as a predictor for A1c reduction and outpatient follow-up [59•]. Education sessions should be short, focused, and spread-out throughout the hospital stay [60•]. Creating a checklist of survival skills can help to ensure that patients have mastered these skills by the time of discharge and are ready to assume basic diabetes self-management at home [40]. The discharge process should also include a mechanism for ensuring an appointment with a diabetes educator for comprehensive DSME after hospital discharge when the patient is feeling well and ready to fully engage in the education process [61•].

In the ideal setting, every patient with diabetes, or at least patients admitted with new or poorly controlled diabetes, should receive education from a diabetes educator during hospitalization. Patients who receive education from a diabetes educator have been observed to experience a lower frequency of readmission [62, 63] and improved post-discharge A1c [54, 59•, 63]. Unfortunately, many hospitals do not have a dedicated diabetes educator. Even in hospitals with a diabetes educator, the percentage of hospitalized patients with diabetes exceeds the ability of a single person (or even two or three) to provide adequate education for all patients with diabetes [44, 64].

Bedside (staff) nurses are in an excellent position to provide patient education since they are in and out of the patient’s room multiple times each day to administer medication and perform other nursing care. Since they are near the patient at all times, they can identify teachable moments when the patient is awake and able to mentally and physically participate in educational efforts. Diabetes education delivered by staff nurses who have the support of a diabetes nurse specialist has been shown to significantly reduce 7-day readmission rates [65]. Unfortunately, many hospital nurses report that they have inadequate support, limited opportunities for their own diabetes-specific education, and competing demands that interfere with their ability to provide necessary information to patients [61•, 66,67,68,69,70]. Hospital nurses report low confidence in providing accurate information on current diabetes therapies, feeling that this falls outside their area of expertise [61•].

One role of the diabetes educator is to serve as a resource and role model for other healthcare providers, including staff nurses [59•]. Diabetes educators can train “volunteer” bedside nurses to be diabetes nurse “champions” who lead educational efforts on their individual patient care units, and who in turn mentor the other staff nurses on their unit [58, 64, 69, 71, 72]. Diabetes educators can create care plans and tools that the bedside nurses can utilize for patient education. In one study, Korytkowski and colleagues found increased patient satisfaction and improved glycemic control during hospitalization in patients randomized to receive an individualized structured diabetes education plan created by a certified diabetes educator (CDE) and delivered by the nursing staff (despite the fact that there was no glycemic management intervention) [52•]. In a prospective nonrandomized study, Magee and colleagues evaluated the impact of a survival skills education intervention administered by CDE-trained assistants who administered a diabetes knowledge pre-test and instructed the patient to watch select videos on a DVD that covered topics related to their knowledge deficits [73••]. This intervention resulted in significant knowledge improvement based on post-test results, and an improvement in self-reported medication self-care behavior from baseline to 2 weeks following discharge. In a third study, Krall and colleagues empowered bedside nurses by providing a standardized teaching tool as well as access to brief patient education videos that focused on key self-management survival skills [61•].

These data demonstrate that bedside nurses, with readily available resources, can successfully provide diabetes survival skills education. If possible, other health professionals within the hospital should also be engaged to establish an interdisciplinary approach to diabetes education. Shah and colleagues randomized hospitalized patients with diabetes to a single 30- to 45-min pharmacist-provided counseling session in addition to nurse-provided education or nurse-provided education alone [46]. Patients in the pharmacist group had better medication self-care behavior following discharge, were more likely to keep follow-up appointments, and had a larger reduction in A1c within a 90-day follow-up period. Hardee and colleagues designed an interdisciplinary diabetes care model in a 909-bed tertiary care center [57••]. Within their model, bedside nurses provided key survival skills education, dieticians saw newly diagnosed patients with diabetes (as well as patients referred to them by nurses or medical providers), pharmacists provided education for complex patients and those with learning barriers, and case managers (social workers or nurses) ensured that discharge medications and glucose monitoring supplies were affordable to the patient. For example, a resource document describing community resources was developed for all neighboring counties which include requirements for referrals and type of payment accepted. These authors achieved cost savings without an increase in LOS or diabetes readmission rates compared to when the hospital had a team of five dedicated diabetes educators (who were only able to see a fraction of patients with diabetes).

In another report, Kimmel and colleagues found that nearly 100% of patients in their hospital received survival skills training when they assessed patient knowledge in a telephone call survey administered 7–10 days following discharge [30]. At their institution, staff nurses teach patients using diabetes teaching folders that are available in four different languages. A diabetes clinical nurse specialist provides continuing education for all nurses and sees patients who have learning difficulties only. A dietician instructs patients on meal planning, and a pharmacist completes a discharge insulin instruction card with the patient.

Ensuring Continuation of Diabetes Care and Education Following Hospital Discharge

Hospitalization has the potential to improve the long-term management of patients with T2DM by providing the opportunity to intensify treatment regimens and provide diabetes education and care coordination [74]. However, despite an intervention including inpatient diabetes management directed by an endocrinologist and intensive inpatient diabetes education from a nurse practitioner CDE, Wexler and colleagues found the observed improvement in A1c from baseline declined from −2.6% at 6 months post-discharge to −1.1% at 12 months post-discharge [54]. This waning of impact highlights the importance of continuity of care and diabetes regimen adjustments in the post-hospitalization time period [75].

Current guidelines recommend outpatient follow-up with the primary care provider (PCP), endocrinologist, or diabetes educator within 30 days of hospital discharge [2, 4, 35], but this may not be soon enough to prevent a diabetes-related readmission. Seggelke and colleagues conducted a pilot study randomizing 100 patients with diabetes, who lacked medical insurance or a PCP, into either a control or an intervention group upon discharge from the hospital [76•]. Patients in the intervention group were seen in a transitional care clinic within 2 to 5 days of hospital discharge. Patients originally admitted for a diabetes-related problem who had the earlier follow-up had fewer readmissions at 90 days (13 vs. 43%).

A goal of successful transition from hospital to home is prevention of hospital readmission and emergency department visits. Multimodal interventions that include monitoring and managing symptoms after discharge (via telephone follow-up or use of transitional care clinics) and providing patients with comprehensive DSME and social support are most likely to be effective in reducing readmissions [77, 78]. There are only a few published studies describing outcomes related to diabetes-specific transition of care programs, and they each used a multimodal approach including intensive diabetes education either prior to or following discharge, and telephone follow-up multiple times during the immediate post-discharge period [59•, 79, 80••]. Each of these studies found significant reductions of A1c for patients receiving the intervention compared to patients receiving usual care. In a 2016 study evaluating the impact of a diabetes transitional care program in a Veterans Affairs (VA) medical center, a nurse practitioner (NP) visited a convenience sample of 40 patients with uncontrolled diabetes (A1c >9%) in the hospital prior to discharge to educate them on survival skills and then provided continuity of care services until the patients had a follow-up appointment with their physician [80••]. Specifically, the NP called the patients weekly, starting 48–72 h after discharge, asked about symptoms, assessed medication adherence, and adjusted diabetes therapy based on individualized care plans. Patients were instructed to call the NP or a nurse care hotline (available 24 h/day, 7 days a week) if they had any problems. Patients who participated in the program had a mean A1c reduction from 11.3% at baseline to 9.1% when measured 2 to 8 months after time of enrollment. The 30-day readmission rate for these high risk patients was 10% which was less than the 14.3% overall readmission rate for all diabetes patients at that institution.

Conclusions

In summary, a patient-centered approach to optimize diabetes medications at discharge, anticipate and manage barriers to post-discharge success, provide survival skills education, and ensure continuation of diabetes care and education has the potential to improve glycemic control and reduce hospital readmissions for many patients with T2DM. The key components for transitioning patients with T2DM from hospital to home include the following:

  1. 1.

    Obtain A1c at time of admission (if not available from the past 2–3 months). Patients with poorly controlled diabetes require intensification of the outpatient regimen at hospital discharge.

  2. 2.

    Perform baseline assessment of patient’s cognitive ability, vision, manual dexterity, current knowledge, numeracy skills, and family support early in the hospitalization.

  3. 3.

    Educate patient and/or patient’s family throughout the hospitalization on basic diabetes survival skills, including insulin administration (if applicable). Education should be tailored to the learning needs identified on baseline and ongoing assessments.

  4. 4.

    Work with patient to identify and manage any potential barriers to a successful transition to home, including financial constraints, lack of an outpatient diabetes provider, transportation issues that impact attendance at follow-up appointments, and medication access issues.

  5. 5.

    Consider patient-specific factors when selecting a discharge diabetes treatment regimen. In patients without a history of diabetic ketoacidosis (DKA) or new symptomatic hyperglycemia, consider starting or intensifying noninsulin options or basal insulin, rather than starting basal-bolus insulin, to lower the risk of hypoglycemia immediately after discharge.

  6. 6.

    Provide patient with all supplies needed to self-manage diabetes at home, including medications (and pen needles or syringes if on insulin) as well as a blood glucose meter, test strips, and lancets. If prescriptions will be given for the patient to fill after discharge, ensure insurance coverage and/or affordability prior to the patient leaving the hospital.

  7. 7.

    Arrange follow-up with the outpatient diabetes provider. If new to insulin, early follow-up (within a few days of discharge) may stave off an ED visit or readmission.

  8. 8.

    Arrange appointment with a CDE for comprehensive DSME whenever possible.

  9. 9.

    Provide patient with clearly written personalized discharge instructions, in a language that they are comfortable reading, that includes a list of all diabetes medications and follow-up appointments. The instructions should also include a contact person for questions or issues, if the patient does not have an established diabetes provider.

  10. 10.

    Communicate the revised diabetes-related treatment plan to outpatient providers.