Original Article

Journal of Pharmacy Practice

1-6
Readmission Outcomes of Sliding Scale ª The Author(s) 2020
Article reuse guidelines:
sagepub.com/journals-permissions
Insulin Compared to Basal-Bolus Insulin DOI: 10.1177/0897190020921028
journals.sagepub.com/home/jpp
Prescribed at Discharge in an Insulin-Naive
Patient Population

Pamela Carter, PharmD, BCACP1 , Tracie Eshelbrenner, PharmD, BCPS1,

Lauren Kirk, PharmD, BCACP1, Mandy Fisk, MPH2 ,
and Claire Rodrigues, PharmD, BCACP1

Abstract
Background: Limited data are available that examine hospital readmission outcomes of sliding scale compared to basal-bolus
insulin in indigent and insulin-naive patients. Objective: To evaluate hospital readmission outcomes in patients who are insulin
naive with type 2 diabetes mellitus who are initiated on either sliding scale or basal-bolus insulin upon hospital discharge.
Methods: A retrospective chart review was conducted of adult patients with a history of type 2 diabetes mellitus, who were
insulin naive, had a hemoglobin A1c (HbA1c) 10% or greater, and were discharged with a prescription for sliding scale or basal-
bolus insulin from January 2015 to July 2018. The primary objective measured all-cause 30-day hospital readmissions. The sec-
ondary objectives measured diabetes-related 30-day hospital readmissions and HbA1c change after 3 months of initial hospital
admission. Data were analyzed using descriptive statistics, w2 test, paired sample t test, and logistic regression. Results: Forty-one
patients were prescribed sliding scale insulin and 105 patients were prescribed basal-bolus insulin. The majority were male (60%),
spoke English (84%), were self-pay (39%), and had a mean age of 51 + 10.2 years, initial HbA1c of 13% + 1.9%, and LACEþ score
of 51 + 15.6 upon discharge. All-cause 30-day hospital readmissions occurred in 14.6% of sliding scale and 6.7% of basal-bolus
insulin groups (odds ratio [OR]: 2.40, 95% confidence interval [CI]: 0.75-7.63). Hyperglycemia occurred in 7.3% of sliding scale
and 0.9% of basal-bolus insulin groups. Mean HbA1c difference for basal-bolus and sliding scale insulin was 3.3 + 3.1 and 2.9 +
2.7, respectively (P ¼ 0.459). Conclusion: There was no significant difference in all-cause 30-day hospital readmissions com-
paring sliding scale to basal-bolus insulin.

Keywords
type 2 diabetes mellitus, hospital readmission, sliding scale insulin, insulin naive, transitions of care

Introduction medications, history of prior hospitalizations, comorbidities,

lower socioeconomic status, poor health literacy, and certain
Hospital readmission is a high priority health-care quality mea-
racial or ethnic minority groups.1
sure, and its prevention is crucial.1 Health-related and social
Once admitted, patients diagnosed with uncontrolled type 2
factors can contribute to an increased risk of hospital readmis-
diabetes may be at an increased risk of complications and
sion. These can include heart failure, myocardial infarction,
mortality due to having consistent hyperglycemia.6 Overall,
pneumonia, chronic obstructive pulmonary disease, and dia-
these patients are initiated on insulin if glycosylated hemoglo-
betes mellitus.2 Diabetes mellitus is one of the most common
bin A1c (HbA1c) is 10% or greater in order to lower their HbA1c
chronic conditions among adults, and its prevalence continues
and prevent future diabetes complications from occurring.7
to increase due to an aging population, changing racial compo-
sition, and rising disease incidence.3,4 About 9.3% of the US
population has diabetes and 28% are undiagnosed.5 This 1
Pharmacy Clinical Services Outpatient, JPS Health Network, Fort Worth,
increasing prevalence can lead to frequent hospital admissions TX, USA
2
and readmissions due to hyperglycemia or hypoglycemia. Peo- Office of Clinical Research, JPS Health Network, Fort Worth, TX, USA
ple with diabetes are more likely to be hospitalized compared
Corresponding Author:
to those without diabetes mellitus, having a diabetes 30-day Pamela Carter, Pharmacy Clinical Services Outpatient, JPS Health Network,
related readmission rate between 14% and 20%. 1 This Fort Worth, TX, USA.
increased risk could be due to patients not being on appropriate Email: [email protected]
2 Journal of Pharmacy Practice XX(X)

Furthermore, once the HbA1c is 10% or greater, antidiabetic also a rarity of data surrounding readmission outcomes in
oral medications alone may not be sufficient to obtain a goal patients with poor health literacy while using a sliding scale
HbA1c of 7% or less.7 Hospitalized patients with diabetes insulin regimen. This information could be beneficial in areas
regardless of HbA1c could be treated with insulin during admis- where indigent patients are cared for including county hospitals
sion due to various factors such as institutional formularies and and facilities. More robust data are warranted to determine the
limitations to oral diabetes medications in regard to patient relationship between these 2 different insulin regimens and
characteristics. Often patients with prior to admission oral dia- readmission outcomes.
betes medications are restarted on the oral medications at hos- At the study institution, prescribers do not have any pre-
pital discharge. However, hospitalized patients with elevated scription restriction on insulin orders and there is no institu-
HbA1c 10% or greater in accordance with the American Dia- tional policy that supports the use of sliding scale insulin at
betes Association may be discharged home on insulin therapy hospital discharge. Regardless of the evidence against sliding
in order to lower their HbA1c.7 scale insulin use, providers at the study institution continue to
Sliding scale and basal-bolus insulin regimens are 2 differ- prescribe sliding scale insulin. The study institution is a county
ent insulin treatment options for these patients. For this study, hospital that treats a large portion of the underserved patient
sliding scale insulin is defined as a rapid-acting insulin (ie, population. The complexity of a sliding scale regimen may not
insulin aspart or insulin lispro) with directions that contain a be appropriate for indigent patients who may also have poor
range for the dose of units or the term “sliding scale,” while health literacy. Due to the paucity in the current literature, this
basal-bolus insulin is defined as having a rapid-acting insulin study would provide insight on readmission outcomes in type 2
that contains a fixed dose of units.7 Both of these regimens can diabetes patients who are initiated on different insulin regi-
contain a basal insulin which is defined as a long-acting insulin mens. The objective of this study is to evaluate hospital read-
(ie, insulin glargine or insulin detemir) that contains a fixed mission outcomes in patients who are insulin naive with type 2
dose of units.7 Basal-bolus insulin is a preferred insulin regi- diabetes who are initiated on either sliding scale or basal-bolus
men since it closely mimics physiological insulin secretion insulin at discharge.
when compared to the sliding scale insulin regimen.6 Many
studies have justified basal-bolus insulin over sliding scale
insulin due to its efficacy in glycemic control.6,8,9 A rando-
Methods
mized control trial compared the efficacy and safety of basal- This was an observational, retrospective chart review of
bolus insulin to sliding scale insulin in insulin-naive patients patients between January 2015 and July 2018. Data were col-
with type 2 diabetes mellitus during inpatient management. lected from the electronic medical record. A list was generated
The authors found that 66% of the basal-bolus insulin group that contained all patients taking either sliding scale or basal-
achieved a blood glucose target of less than 140 mg/dL, while bolus insulin who were newly started on insulin upon discharge
only 38% in the sliding scale insulin group achieved the target during the specified time frame. There was a small amount of
(P < 0.01).6 A retrospective study found basal-bolus insulin patients on sliding scale insulin compared to basal-bolus insu-
attained lower fasting blood glucose levels (194.4 vs lin; therefore, all patients were reviewed who were on sliding
208.9 mg/dL; P ¼ 0.028) and mean blood glucose levels scale insulin. For patients on basal-bolus insulin, a random
(221.4 vs 230.4 mg/dL; P ¼ 0.021) compared to sliding scale sample of patients were included using a random number gen-
insulin regimens in severe hyperglycemia. 8 Therefore, the erator due to the high volume. This study received approval
American Diabetes Association strongly discourages the sole from the North Texas Regional Institutional Review Board on
use of sliding scale insulin due to the lack of glycemic control.7 October 16, 2018.
Regardless of the evidence for use of sliding scale insulin, Adults with type 2 diabetes were included if they met all of
sliding scale insulin is being prescribed upon discharge. If a the following criteria: 18 years and older, insulin naive, HbA1c
patient is newly started on insulin with uncontrolled type 2 10% or greater, discharged from inpatient or observation units,
diabetes mellitus, they could be at an increased risk of read- and discharged home with a prescription for sliding scale insu-
mission if discharged on sliding scale insulin. The complexity lin or basal-bolus insulin. Sliding scale insulin was defined as a
of a sliding scale insulin regimen could potentially lead to poor medication order for insulin aspart or insulin lispro with direc-
patient outcomes including increased readmissions, adverse tions that contained a range for the dose of units or the term
drug events, and inappropriate glycemic control.10 Currently, “sliding scale.” Basal-bolus insulin was defined as a medica-
there are limited data available that examine the readmission tion order for insulin aspart or insulin lispro with directions that
outcomes of sliding scale insulin compared to basal-bolus insu- contained a fixed dose of units. Both regimens could contain
lin in this specific patient population. A prospective study did basal or long-acting insulin including insulin glargine or insulin
find that the use of sliding scale insulin in older patients was detemir. Adults were excluded from the study if they had a
associated with an increased risk of hospitalization (odds ratio history of type 1 diabetes, were pregnant, discharged from the
[OR]: 4.97).11 However, a randomized controlled trial in a emergency department or urgent care, had a primary care pro-
long-term care facility found sliding scale insulin and basal- vider outside of the health system, discharged on an oral corti-
bolus insulin had similar rates of hypoglycemia, hyperglyce- costeroid or sulfonylurea, or were prescribed Human NPH
mia, and hospitalizations in nursing home residents.12 There is insulin or premixed insulins. Since the study institution cares
Carter et al 3

for the uninsured and underinsured patient population, it

receives 340b pricing which allows for more affordable insulin 395 Paents Screened
glargine, insulin detemir, insulin aspart, and insulin lispro. In
addition, insulin formulations outside of long-acting and rapid- 249 Paents Excluded:
acting insulins were excluded to focus purely on the effects of Insulin experienced: 168
sliding scale of the rapid-acting insulin compared to basal- Type 1 Diabetes Mellitus: 29
bolus insulin on readmission. Prescribed sulfonylurea: 17
Prescribed NPH or premixed insulin: 13
In addition to patient demographics, data points that were
Primary care provider outside of instuon: 11
collected were if the patient was admitted to the intensive care Prescribed oral corcosteroids: 8
unit during hospitalization, had an established primary care Pregnancy: 3
provider, had a hospital discharge appointment scheduled,
attended the hospital discharge appointment, had their insulin 146 Paents Included
dose adjusted after discharge, and received diabetes education
from a nurse educator before discharge. Also, the LACEþ
41 Paents with Sliding Scale 105 Paents with Basal-Bolus
score was included for each patient upon discharge. The Insulin Insulin
LACEþ index was utilized to determine the patient’s hospital
readmission risk. The LACEþ index calculates patients risk for
hospital readmission or death within 30 days of discharge, and Figure 1. Patient enrollment.
a score greater than 58 was considered high risk at the study
institution.13 It was also documented if a patient received a analyze the all-cause 30-day readmission and Fisher’s exact
phone call from a transition of care (TOC) clinical pharmacist test was used to analyze the diabetes-related 30-day readmis-
after discharge. There was a TOC pharmacist at the institution sions. Paired sample t test was used to analyze the mean out-
who would receive consults based on certain criteria to pro- come of the change in HbA1c percentage after 3 months of
vide clinical services such as one-on-one time with the patient initial hospital admission. A bivariate logistic regression was
to review medications, answer patient questions, clarify dis- used to evaluate the variables associated with the all-cause and
crepancies, and provide medication and disease state educa- diabetes-related 30-day readmission outcomes. A simple linear
tion. The TOC pharmacist consult criteria included patients regression model was used to evaluate the variables associated
being discharged home with at least one of the following: with a change in HbA1c. Statistical significance was deter-
newly prescribed insulin, diabetes-related admission (diabetic mined with an alpha value being equal to or less than 0.05.
ketoacidosis [DKA], diabetes foot ulcer, hyperglycemia, or Statistics were performed using SAS® 9.4 (SAS Institute, Cary,
hypoglycemia), or likelihood of low medication adherence. North Carolina).
A TOC pharmacist would then conduct phone calls within
2 days of discharge to provide a medication reconciliation
and identify any medication-related problems. They would
Results
also have the ability to adjust the patients’ insulin regimen Of the 395 patients screened, 249 patients were excluded and
according to an approved institutional protocol based on 146 patients were included for analysis (Figure 1). There were
patient-reported blood glucose levels. In addition to insulin 41 patients discharged on sliding scale insulin and 105 patients
dose adjustments, this protocol allows the TOC pharmacist to discharged on basal-bolus insulin. The baseline characteristics
change the sliding scale insulin regimen to a basal-bolus insu- are listed in Table 1.
lin regimen. The TOC pharmacist would stop attempting to The majority of patients were male (60%), 51 + 10.2 years
reach the patient after 3 phone call attempts were made. old, non-Hispanic (66%), had a body mass index of 32 + 8-
Patients could have been contacted by the TOC pharmacist kg/m2, spoke English (84%), and did not have any insurance or
in either insulin group. funding source, also known as self-pay (39%). The average
The primary outcome measured the all-cause 30-day read- initial HbA1c for sliding scale insulin and basal-bolus insulin
mission. The secondary outcomes assessed the diabetes-related groups was 12.9% + 2% and 12.9% + 1.9%, respectively. The
30-day readmission defined as hypoglycemia, DKA, hyperos- number of emergency department/urgent care visits and hospital
molar hyperglycemic state (HHS), or hyperglycemia. Hypogly- admissions 1 year prior to hospitalization averaged 1.2 + 1.7
cemia was defined as blood glucose less than 70 mg/dL and and 0.6 + 1.4, respectively. Blood glucose on initial admission
hyperglycemia was defined as the first blood glucose greater and upon discharge were noted to be 420 + 156 mg/dL and 223
than 300 mg/dL collected on arrival at readmission excluding + 66 mg/dL, respectively. Only 6% of patients were admitted to
DKA and HHS. The other secondary outcomes assessed the the intensive care unit during the initial hospital admission.
change in HbA1c percentage after 3 months of initial hospital Forty-eight percent of patients were established with a primary
admission. care provider before the initial hospital admission. The average
Multiple statistical procedures were utilized for data analy- LACEþ score for patients was 51 + 16 on discharge. A hospital
sis. Descriptive statistics were used to describe the basic fea- discharge appointment was scheduled in 92% of patients and
tures of the study population. Yates’ chi-square test was used to 69% of patients attended that appointment. Twenty-three percent
4 Journal of Pharmacy Practice XX(X)

Table 1. Baseline Characteristics.

Characteristic Overall (N ¼ 146) Sliding Scale (n ¼ 41) Basal-Bolus (n ¼ 105) P Value

Mean + SD age, years 50.9 + 10.2 49.7 + 10.9 51.3 + 9.9 0.387
Mean + SD BMI, kg/m2 32.1 + 7.9 31.2 + 7.3 32.5 + 8.2 0.354
Sex, no. (%) 0.362
Male 87 (59.6) 22 (53.7) 65 (61.9)
Female 59 (40.4) 19 (46.3) 40 (38.1)
Ethnicity, no. (%) 0.143
Non-Hispanic 97 (66.4) 31 (75.6) 66 (62.9)
Hispanic 49 (33.6) 10 (24.4) 39 (37.1)
English speaking, no. (%) 122 (83.6) 36 (87.8) 86 (81.9) 0.464
Insurance, no. (%) 0.022
Self-pay 57 (39.1) 19 (46.3) 38 (36.2)
Institution-funded insurance 40 (27.4) 12 (29.3) 28 (26.7)
Government 34 (23.3) 3 (7.3) 31 (29.5)
Commercial 15 (10.3) 7 (17.1) 8 (7.6)
Mean + SD initial HbA1c, % 13.0 + 1.9 12.9 + 1.9 13.0 + 1.9 0.641
Mean + SD number of ED or UC visitsa 1.2 + 1.7 1.0 + 1.5 1.3 + 1.8 0.286
Mean + SD number of admissionsa 0.6 + 1.4 0.6 + 1.2 0.7 + 1.4 0.781
Mean + SD blood glucose on admission, mg/dL 419.5 + 156.1 449.9 + 201.1 407.6 + 133.9 0.490
Mean + SD blood glucose at discharge, mg/dL 223.5 + 65.8 224.2 + 76.2 223.2 + 61.6 0.934
Admitted to ICU during hospitalization, no. (%) 9 (6.2) 5 (12.2) 4 (3.8) 0.641
Established primary care provider, no. (%) 70 (48.0) 17 (41.5) 53 (50.5) 0.327
Mean + SD LACEþ score on discharge 51.1 + 15.6 53.5 + 15.6 50.1 + 15.6 0.235
Hospital discharge appointment scheduled, no. (%) 134 (91.8) 36 (87.8) 98 (93.3) 0.274
Attended hospital discharge appointment, no. (%) 101 (69.2) 23 (56.1) 78 (74.3) 0.032
Insulin dose adjusted after discharge, no. (%) 61 (41.8) 9 (22.0) 52 (49.5) 0.016
Received TOC pharmacist phone call after discharge, no. (%) 33 (22.6) 7 (17.1) 26 (24.8) 0.318
Received diabetes education before discharge, no. (%) 31 (21.2) 12 (29.3) 19 (18.1) 0.138
a
One year prior to initial hospitalization.
Abbreviations: BMI, body mass index; ED, emergency department; HbA1c, hemoglobin A1c; ICU, intensive care unit; SD, standard deviation; TOC, transitions of
care; UC, urgent care.

Table 2. Readmission Outcomes.

Overall Sliding Scale Basal-Bolus Odds Ratio

Outcome (N ¼ 146) (n ¼ 41) (n ¼ 105) (95% Confidence Interval)

All-cause 30-day readmission, no. (%) 13 (8.9) 6 (14.6) 7 (6.7) 2.40 (0.75-7.63)
Readmission status, no. (%)
a
Inpatient 9 (6.2) 4 (9.8) 5 (4.8)
Observation 4 (2.7) 2 (4.9) 2 (1.9)
Diabetes-related 30-day readmissions, no. (%)
a
Hyperglycemia 4 (2.7) 3 (7.3) 1 (0.9)
Hyperosmolar hyperglycemic state 0 (0) 0 (0) 0 (0)
Diabetic ketoacidosis 0 (0) 0 (0) 0 (0)
Hypoglycemia 0 (0) 0 (0) 0 (0)
a
Not reported.

of patients received a phone call from a TOC clinical pharmacist The primary outcome evaluating all-cause 30-day readmis-
after discharge, and 42% had their insulin regimen adjusted sions and secondary outcomes evaluating 30-day diabetes-
either during the TOC phone call or at the hospital discharge related readmissions are found in Table 2.
appointment. This insulin regimen adjustment could have been The sliding scale insulin group had 6 patients (14.6%) and
the patient being switched from sliding scale insulin to basal- basal-bolus insulin group had 7 patients (6.7%) readmitted
bolus insulin or having their insulin dose adjusted. Twenty-one within 30 days for any cause, having an OR of 2.40 (95%
percent of patients received diabetes education from a nurse confidence interval: 0.75-7.63). Out of the 13 patients who
educator before hospital discharge. readmitted, only 4 patients (31%) had a LACEþ score greater
Carter et al 5

Table 3. Change in Hemoglobin A1c. patients aged 65 years and older found that sliding scale insulin
was associated with an increased risk of rehospitalization; how-
Mean + SD
ever, this study did not include basal-bolus insulin nor com-
Sliding Scale Basal-Bolus P pared sliding scale insulin to any other insulin regimen.11 Our
Outcome (n ¼ 41) (n ¼ 105) Value study was able to compare both insulin groups with regard to
multiple outcomes including safety and efficacy outcomes. We
Initial HbA1c, % 12.9 + 1.9 12.99 + 1.9 .736
HbA1c 3 months after admission, % 10.00 + 2.5 9.71 + 2.8 .574
examined patients who were newly started on insulin to deter-
Difference in HbA1c, % 2.87 + 2.7 3.27 + 3.1 .459 mine if that had an effect on our outcomes. Also, we were able
to include patients who received care at a county hospital where
Abbreviation: SD, standard deviation. most patients have limited health-care funding.
Overall, there were few readmission events. This could have
than 58. Overall, 4 patients meet the criteria for hyperglycemia been due to the fact that our current policies and procedures in
as a diabetes-related 30-day readmissions with 3 patients in the place are effective. This includes having clinical pharmacists
sliding scale group and 1 patient in the basal-bolus group. No (TOC and non-TOC clinical pharmacists) reviewing discharge
patients met the criteria for the hypoglycemia, HHS, or DKA medications prior to discharge. This was an opportunity for
within the diabetes-related 30-day readmission outcome. The patients on sliding scale insulin regimen be switched to
secondary outcomes evaluating change in HbA1c can be found basal-bolus insulin. Another example includes the TOC clinical
in Table 3. pharmacist completing posthospital follow-up phone calls with
There was a mean HbA1c difference of 2.9% + 2.7% in patients. During that time, the patient could have their insulin
sliding scale group and 3.3% + 3% in the basal-bolus group (P regimen changed from a sliding scale regimen to a basal bolus
¼ 0.459). Three variables were identified as being significant regimen. About 40% of all patients had their insulin dose
in the logistic regression model with regard to effect on the adjusted after discharge, which could have had a role in reduc-
primary and secondary outcomes. The 3 variables that were ing the readmission risk. Reasons that there is a low percentage
significant with regard to difference in HbA1c included having of patients in this study who received the TOC pharmacist
a hospital discharge appointment scheduled (P ¼ 0.002), intervention include dependence based on consults ordered,
patients who attended the hospital discharge appointment (P no process for the TOC pharmacist to identify patients dis-
¼ 0.001), and patients who received a clinical pharmacist TOC charged on sliding scale insulin, and whether or not the patient
phone call after discharge (P ¼ 0.001). There was one variable answers the phone call.
that was significant for all-cause 30-day readmission which Having a larger sample would have also helped to detect a
included patients who attended the hospital discharge appoint- difference with regard to the readmission outcomes. Also, the
ment (P ¼ 0.001). sample was not at high risk of readmission with an average
LACEþ score of 51. Most of the patients included did not have
any insurance or funding source, so there is uncertainty if
patients were able to afford their medication or if they were
Discussion not taking their medications at all. For patients newly started on
In this retrospective, observational study evaluating the differ- insulin, it is imperative that they receive education regarding
ence of sliding scale and basal-bolus insulin on hospital read- insulin administration. The diabetes education upon discharge
missions, it was found that there was no statistical difference in for these patients newly started on insulin was inconsistent,
all-cause and diabetes-related 30-day readmissions comparing which could have been due to limited staffing personnel and
the 2 insulin groups. In addition, we found that there was no resources. Patients had the opportunity to get diabetes educa-
statistical difference with regard to difference in HbA1c com- tion from nursing staff, diabetes nurse educators, and TOC
paring sliding scale insulin to basal-bolus insulin, but the basal- pharmacists. With regard to the diabetes nurse educators, there
bolus insulin group did have a greater reduction in HbA1c. This were only 2 who assisted with diabetes education to the patient
was the first study that evaluated outcomes of insulin regimen prior to discharge. They were available Monday through Friday
on hospital readmissions and assessed different variables that during normal business hours to provide education, which
could affect hospital readmissions in an insulin-naive, indigent could have been a reason why they reached a low number of
patient population. patients. If the patient did not receive education by this diabetes
The results from a randomized control trial showed that both nurse educator at discharge, it is possible the patient received
sliding scale and basal-bolus insulin groups did not have a education from the TOC pharmacist either prior to discharge or
difference with regard to hospital readmissions.12 They also over the phone after discharge.
found that basal-bolus insulin had a lower 3-day average fast- This is the first study to evaluate readmissions in a patient
ing blood glucose compared to sliding scale insulin (P ¼ population with poor health literacy and who are insulin naive.
0.023). However, they did not include whether these patients We compared sliding scale insulin and basal-bolus insulin in
were insulin naive nor did they describe socioeconomic factors both safety and efficacy outcomes of hospital readmission and
that could have effected their outcomes. A prospective cohort change in HbA1c in this specific patient population. We were
study that assessed potentially inappropriate prescriptions in able to include a large time frame and assessed the patients over
6 Journal of Pharmacy Practice XX(X)

a 30-day period to appropriately assess the 30-day readmission Funding

outcomes. However, our results should be interpreted with cau- The author(s) received no financial support for the research, author-
tion due to the retrospective study design and small sample ship, and/or publication of this article.
size. We also did not measure adherence of the insulin regi-
mens such as if patients picked up the prescriptions and if ORCID iDs
patients were taking the insulin regimens as instructed. This Pamela Carter, PharmD https://orcid.org/0000-0002-8572-2752
could not be easily measured since patients could have filled Mandy Fisk, MPH https://orcid.org/0000-0003-2045-0327
their medications at an outside pharmacy. A follow-up period
more than 30 days would be helpful to assess in future studies References
to determine the HbA1c reduction and effects of readmission 1. Rubin DJ. Hospital readmission of patients with diabetes. Curr
rates of these 2 regimens in the long term due to diabetes being
Diabet Rep. 2015;15(2):17.
a chronic condition. With regard to the sliding scale dose, we
2. AHRQ Patient Safety Network. Rockville, MD: Agency for
did not gather details for the sliding scale dose used or if the
Healthcare Research and Quality, https://www.ahrq.gov/cpi/
patient was using the sliding scale as prescribed or taking a set
about/otherwebsites/psnet.ahrq.gov/index.html. Accessed Janu-
dose in its place. There are also various outside factors includ-
ary 2020.
ing comorbidities that we did not account for which could have
3. Jiang HJ, Stryer D, Friedman B, et al. Multiple hospitalizations
impacted our outcomes. This could potentially be a reason why
for patients with diabetes. Diabetes Care. 2003;26(5):1421-1426.
this study did not show a statistical difference in the 30-day 4. Hirschman KB, Bixby MB. Transitions in care from the hospital
hospital readmissions outcome. This study was also not pow-
to home for patients with diabetes. Diabetes Spect. 2014;27(3):
ered a priori due to the lack of previous literature surrounding
192-195.
this topic. Lastly, it was not feasible to objectively measure
5. Ostling S, Wyckoff J, Ciarkowski SL, et al. The relationship
health literacy due to the retrospective nature of this study.
between diabetes mellitus and 30-day readmission rates. Clin
This is due to the fact that the study institution does not routi-
Diabet Endocrinol. 2017;3(3):1-8.
nely screen health literacy in the electronic medical record at
6. Umpierrez GE, Smiley D, Zisman A, et al. Randomized study of
this time.
basal-bolus insulin therapy in the inpatient management of
A future direction for the study institution includes creating patients with type 2 diabetes (RABBIT 2 Trial). Diabetes Care.
a process for the TOC pharmacist to identify patients on sliding
2007;30(9):2181-2186.
scale insulin regimens prior to discharge. This process could
7. Riddle MC, Bakris G, Blonde L, et al. Standards of medical care
assist in changing their insulin regimen from a sliding scale
in diabetes—2019. Ame Diabetes Associat. 2019;(42):S1-S193.
insulin regimen to a basal-bolus insulin regimen before dis-
8. Huri HZ, Permalu V, Vethakkan SR. Sliding-scale versus basal-
charge. This could prevent any confusion for the patient on
bolus insulin in the management of severe or acute hyperglycemia
how to use their insulin regimen prior to discharge to prevent
in type 2 diabetes patients: a retrospective study. Plos One. 2014;
hospital readmissions and improve HbA1c reduction.
9(9):1-9.
9. Roberts GW, Agullar-Loza N, Esterman A, et al. Basal-bolus
Conclusion insulin versus sliding scale insulin for inpatient glycaemic con-
In conclusion, basal-bolus insulin had a smaller occurrence of trol: a clinical practice comparison. MJA. 2012;196(1):266-269.
all-cause 30-day and diabetes-related hospital readmissions 10. Jackson B, Grubbs L. Basal-bolus insulin therapy and glycemic
comparing to sliding scale, but it was not statistically signifi- control in adult patients with type 2 diabetes mellitus: a review of
cant. It was also found that basal-bolus insulin had a greater the literature. J Am Assoc Nurse Pract. 2014;26(6):348-352.
reduction in HbA1c, but the difference was not significant. A 11. Bo M, Quaranta V, Fonte G, et al. Prevalence, predictors and
larger, prospective, randomized trial is warranted to obtain clinical impact of potentially inappropriate prescriptions in
conclusive results, as well as, including a uniform way of hospital-discharged older patients: a prospective study. Geriatr
assessing health literacy. Gerontol Int. 2018;18(4):561-568.
12. Dharmarajan TS, Mahajan D, Zambrano A, et al. Sliding scale
Author’s Note insulin vs basal-bolus insulin therapy in long-term care: a 21-day
Lauren Kirk is now affiliated with General Internal Medicine, OSU randomized controlled trial comparing efficacy, safety and feasi-
Physician’s, Inc., Columbus, Ohio, USA. bility. JAMDA. 2016;17(3):206-213.
13. Van Walraven C, Wong J, Forester AJ. LACEþ index: extension
Declaration of Conflicting Interests of a validated index to predict early death or urgent readmission
The author(s) declared no potential conflicts of interest with respect to after hospital discharge using administrative data. Open Med.
the research, authorship, and/or publication of this article. 2012;6(3):e90-e100.