ORIGINAL ARTICLE

E n d o c r i n e C a r e

Clinical Inertia of Discharge Planning among Patients

with Poorly Controlled Diabetes Mellitus

Michelle L. Griffith, Jeffrey B. Boord, Svetlana K. Eden, and Michael E. Matheny

Center for Health Services Research (J.B.B.) and Geriatric Research, Education, and Clinical Care Center
(S.K.E., M.E.M.), Veterans Affairs Tennessee Valley Health System, Nashville, Tennessee 37212; Division
of Diabetes, Endocrinology, and Metabolism (M.L.G.), Department of Medicine, and Departments of
Biostatistics (M.E.M.), Medicine (M.E.M.), and Biomedical Informatics (M.E.M.), Vanderbilt University
School of Medicine, Division of Endocrinology and Metabolism (M.L.G.), Department of Medicine,
University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania 15260; Vanderbilt Heart and
Vascular Institute (J.B.B.), Nashville, Tennessee 37232

Objective: We examined the effect of hospital admissions on the medical treatment of poorly
controlled diabetes mellitus among Veterans Affairs (VA) patients.

Research Design and Methods: This retrospective cohort study included male patients admitted to
one of three VA hospitals from July 1, 2002, to August 31, 2009, who were receiving medication
therapy for diabetes with hemoglobin A1c (HgbA1c) greater than 8.0%. The primary outcome was
a change in preadmission and outpatient prescriptions for diabetes at hospital discharge. Cova-
riates for multivariable logistic regression analysis of the primary outcome were defined a priori
and retrieved from the electronic health record.

Results: Of 2025 admissions for 1359 patients, 454 had some change in diabetes medications at
discharge (rate of change 22.4%). In an adjusted analysis, higher preadmission HgbA1c [odds ratio
(OR) 1.12 per 1.0 U increase; 95% confidence interval (CI) 1.12–1.05; P ⬍ 0.001], higher mean blood
glucose during admission (OR 1.07 per 10 mg/dl increase; 95% CI 1.05–1.10; P ⬍ 0.0001), occurrence
of inpatient hypoglycemia (blood glucose ⬍ 50 mg/dl; OR 1.82, 95% CI 1.32–2.51, P ⬍ 0.001), and
inpatient basal insulin therapy (OR 1.71; 95% CI 1.25–2.35; P ⬍ 0.001) were associated with higher
odds of change in therapy. A total of 656 admissions (32%) demonstrated aggregate clinical inertia
with no change in therapy, no documentation of HgbA1c within 60 d of discharge, and no fol-
low-up appointment within 30 d of discharge.

Conclusions: In this multicenter, retrospective study of patients with poorly controlled diabetes
and at least one hospitalization, less than a quarter received a change in outpatient diabetes
therapy upon discharge, suggesting widespread clinical inertia. Nearly one third had no change in
therapy or subsequent follow-up scheduled. (J Clin Endocrinol Metab 97: 2019 –2026, 2012)

iabetes mellitus is a chronic disease with a significant
D impact on morbidity and mortality and high prev-
alence among the population of U.S. veterans (1). Poor
glycemic control is an important factor in the development
of microvascular complications, such as retinopathy, ne-
phropathy, and neuropathy. It is also associated with in-
creased long-term cardiovascular risk and negatively im-
pacts wound healing and infection risk (2– 4).
Clinical inertia, defined as failure to initiate or intensify
therapy when it is clinically indicated, has been docu-
mented in the care of diabetes in outpatient settings, within
the Veterans Affairs (VA) health system and other systems
(5–9). In a study of outpatient diabetes treatment at several
VA hospitals over 16 months, treatment intensification
occurred in less than 9.8% of visits, despite 39% of pa-
tients having hemoglobin A1c (HgbA1c) greater than 8%

ISSN Print 0021-972X ISSN Online 1945-7197 Abbreviations: CI, Confidence interval; HgbA1c, hemoglobin A1c; IQR, interquartile range;
Printed in U.S.A. NPH, neutral protamine Hagedorn; OR, odds ratio; VA, Veterans Affairs.
Copyright © 2012 by The Endocrine Society
doi: 10.1210/jc.2011-3216 Received November 23, 2011. Accepted March 2, 2012.
First Published Online March 30, 2012

J Clin Endocrinol Metab, June 2012, 97(6):2019 –2026 jcem.endojournals.org 2019

2020 Griffith et al. Clinical Inertia of DM Discharge Planning
at entry to the study (9). Clinical inertia for diabetes treat-
ment has also been documented in inpatient settings (10 –
12). A study of 3613 hospitalized general ward patients
with diabetes demonstrated intensification of antihyper-
glycemic medications in 22% of days with hyperglycemia.
When treatment was intensified, it was associated with a
decrease in daily mean glucose, but hypoglycemia (blood
glucose ⬍ 50 mg/dl) was uncommon; hypoglycemia was
observed in 2.2% of hyperglycemic patients on the day
after treatment intensification and occurred in similar
rates in groups with and without a prior treatment inten-
sification (10).
However, little work has been published on the in-
patient-to-outpatient transition for diabetes care or on
diabetes-specific discharge planning and postdischarge
follow-up (13). Admissions to the hospital represent an
important opportunity to improve glycemic control
among poorly controlled diabetes patients by adjusting
outpatient medications at the point of discharge and
providing appropriate outpatient clinical follow-up. To
our knowledge, no one has examined provider behavior
at discharge in a large cohort of patients with poorly
controlled diabetes. We evaluated the impact of admis-
sion events among four VA hospitals on the outpatient
treatment of diabetes.
Materials and Methods
Study population, setting, and design
This retrospective cohort study included VA patient admis-
sions in Louisville, KY, Huntington, WV, and the Tennessee
Valley Healthcare System, which includes integrated hospitals in
Nashville and Murfreesboro, TN, from July 1, 2002, to August
31, 2009. The VA has a fully electronic health record that in-
cludes a laboratory test system, radiology, bar-coded medication
administration, integrated outpatient and inpatient pharmacy
systems, administrative and billing, and text documents. To
identify patients with poor diabetes control, patient admissions
were eligible for inclusion if an HgbA1c of greater than 8.0% was
measured within 6 months before admission. A record of at least
one filled diabetes medication prescription was required to be
filled within the VA system in the 4 months before admission to
exclude patients who may have received diabetes treatment ex-
clusively by providers outside the VA health system. Records
were excluded if patients were deceased at hospital discharge or
were coded as receiving palliative care during the hospitalization
(International Statistical Classification of Diseases and Related
Health Problems, ninth revision, code V66.7). Female patient
admissions were excluded because they represented only 3% of
the eligible population in this study (65 admissions, 48 subjects),
and this was an inadequate sample size to draw inference from
the results of the multivariable analysis for this subpopulation.
Veterans Affairs Tennessee Valley Healthcare System Institu-
tional Review Board and Research and Development Committee
approvals were obtained and a Veterans Integrated Service Net-
J Clin Endocrinol Metab, June 2012, 97(6):2019 –2026
work (VISN) 9 Data Warehouse data use agreement was in place.
The research was approved as an expedited protocol for retrospec-
tive chart review with waived informed consent and Health Insur-
ance Portability and Accountability Act of 1996 notification.
Outcome definitions
The primary outcome of interest was a change in preadmis-
sion, outpatient prescriptions for diabetes medications at hos-
pital discharge. Prescriptions filled during the 4-month interval
before admission were used to determine baseline diabetes ther-
apy. Prescriptions written at discharge (time stamped within 24 h
before or after discharge) signified addition of medications or
alteration of dosing at discharge; preexisting medications refilled
within 4 months of discharge were used to determine continu-
ation vs. discontinuation of a preadmission medication. Total
daily doses of individual agents were compared on a per-drug
basis, at preadmission, and at discharge or after discharge. We
evaluated any change in therapy as an indicator of attention to
diabetes therapy rather than intensification alone because in-
stances such as patient nonadherence or history of significant
outpatient hypoglycemia may have made a decision to decrease
diabetes therapy appropriate. The primary outcome included an
increase or decrease in dose for an existing diabetes medication,
addition of a medication, discontinuation of a medication, or
change in dose or type of insulin (for example, change of an
neutral protamine Hagedorn (NPH)-based regimen to glargine).
Data collection
All data were retrieved from the VA electronic health record.
Variables that had a clinically plausible impact on providers’
decisions about outpatient diabetes therapy at the point of dis-
charge were collected, including most recent HgbA1c (either pre-
ceding the admission or collected during the admission), previ-
ous admission to these VA hospitals within the previous 90 d,
inpatient glucose measurements, preadmission and inpatient se-
rum creatinine measurements, a diagnosis of chronic kidney dis-
ease (International Statistical Classification of Diseases and Re-
lated Health Problems, ninth revision, codes 585, 586, and
588.8), inpatient diabetes medications including use of a basal
insulin (glargine, detemir, NPH, or a mixed insulin with NPH
component), and presence of an inpatient endocrinology con-
sultation request. Follow-up parameters were collected within
the same hospital system, including outpatient pharmacy orders
and fill records, outpatient primary care or endocrinology ap-
pointments within 30 d of discharge, and HgbA1c values within
60 d of discharge. Thirty-day follow-up is consistent with Amer-
ican Diabetes Association standards of care for patients with
inpatient hyperglycemia (14). Relatively short-term reassess-
ment of HgbA1c may be performed in situations in which a
patient had recently made a change expected to have an impact
on diabetes care. Aggregate clinical inertia was defined as failure
to make a change in therapy, schedule follow-up within 30 d, or
obtain HgbA1c within 60 d.
Statistical analysis
Patient characteristics were summarized using median and
interquartile range (IQR) for continuous variables and frequency
(percentage) for categorical variables. Univariate and multivari-
able analyses for the primary outcome were performed using
logistic regression with Huber-White robust SE estimator to ac-
count for intraclass correlation resulting from multiple admis-
J Clin Endocrinol Metab, June 2012, 97(6):2019 –2026 jcem.endojournals.org 2021

sions per patient (15–17). The multivariable model included co-

TABLE 1. Patient admission characteristics
variates with a clinically plausible relationship to the outcome,
defined a priori: age, race, hospital site, most recent HgbA1c, Variable Total cohort
maximum inpatient serum creatinine, last inpatient serum cre- Total patient admissions 2025
atinine, mean inpatient glucose, last inpatient glucose, admis- Unique patients 1359
sion service (medical vs. surgical/mixed/other), days from Male 2025 (100%)
start of the study period to patient admission date, Centers for Age (yr) 60 (54, 68)
Medicare and Medicaid Services major diagnostic category, Site
binary variables for hypoglycemia during admission (⬍50 mg/ Huntington 462 (23%)
dl), inpatient basal insulin therapy, previous admission within Louisville 585 (29%)
90 d, inpatient endocrinology consultation, and length of stay. Tennessee Valley Healthcare System 978 (48%)
Race
The length of stay was adjusted with the mean Centers for
White 1576 (80%)
Medicare and Medicaid Services length of stay for each diag- African-American 379 (19%)
nosis-related group. For multivariable analysis, missing vari- Other 18 (1%)
ables were imputed using multiple imputation (18, 19). Find- Length of stay (d) 3 (1, 6)
ings with a two-sided P ⱕ 0.05 were considered statistically Most recent HgbA1c (%) 9.6 (8.7, 11.1)
significant. Statistical analyses were performed in statistical Days from most recent HgbA1c 89 (35, 134)
language R, version 2.12.1 (http://www.r-project.org/). to admission
Any insulin prescription filled within 1225 (60%)
4 months prior to admission
Major diagnostic category
Results Nervous system 237 (12%)
Eye disorders 116 (6%)
Cohort description Ear, nose, mouth, and throat 150 (7%)
Respiratory system 173 (9%)
The cohort included 2025 admissions for 1359 unique
Circulatory system 634 (31%)
male patients. The median most recent HgbA1c was 9.6% Digestive system 75 (4%)
(IQR 8.7–11.1%), at a median of 89 d (IQR 35–134) Hepatobiliary system and pancreas 63 (3%)
before admission. Approximately 2% of admissions (40 of Musculoskeletal/connective tissue 164 (8%)
Skin, sc tissue and breast 70 (3%)
2025) had a diagnosis-rated group within the major diag- Endocrine/nutritional/metabolic 40 (2%)
nostic category including diabetes. Other characteristics are Kidney and urinary tract 47 (2%)
listed in Table 1. Adjusted length of stay, maximum inpatient Mental diseases and disorders 96 (5%)
Alcohol/drug use and alcohol/ 75 (4%)
serum creatinine, last inpatient serum creatinine, last inpa- drug-induced organic mental disorders
tient glucose, mean inpatient glucose, and inpatient hypo- Other 85 (4%)
glycemia had missing instances with a maximum of instances Admission service
Medical 1385 (68%)
missing for a single variable of 12.5%. Surgical 292 (14%)
Of 2025 admissions, 454 had some change in diabetes Mixed (telemetry) 194 (10%)
medications at discharge, giving a rate of change of Other (psychiatry, rehabilitation) 154 (8%)
22.4%. A summary of the admission characteristics ac- Categorical variables are presented as counts with percentages, and
cording to whether the patient received any change in di- continuous variables are presented as median with IQR (25th
percentile, 75th percentile).
abetic therapy at discharge are shown in Table 2. Individ-
ual variables that were associated with change in therapy
are shown in Table 3. An increase in preadmission insulin therapy, 95% CI 1.2–1.9; P ⬍ 0.001), any insulin
HgbA1c by 1 U was associated with a 19% increase in therapy during admission (OR 1.97; relative to no insulin
odds of having change in therapy [odds ratio (OR) 1.19 therapy during the admission, 95% CI 1.23–3.18; P ⬍
per 1 U increase; 95% confidence interval (CI) 1.12–1.26; 0.005), or were receiving inpatient basal insulin therapy
P ⬍ 0.0001]. In addition, higher mean blood glucose dur- (OR 2.6; relative to no basal insulin therapy, 95% CI
ing admission was associated with a 6% increase in odds 1.9 –3.4; P ⬍ 0.0001). Being African-American increased
of having change in therapy (OR 1.06 per 10 mg/dl in- odds of having a change in therapy (OR 1.49 relative to
crease; 95% CI 1.04 –1.08; P ⬍ 0.0001). Occurrence of white race; 95% CI 1.13–1.96; P ⬍ 0.005).
hypoglycemia with blood glucose less than 50 mg/dl dur- Table 4 summarizes the results of the adjusted analysis.
ing the admission increased the odds of the outcome by After adjustment for other patient characteristics, a 1-U
84% (OR 1.84, compared with patients with blood glu- increase in patient’s HgbA1c was associated with a 12%
cose 50 mg/dl or greater, 95% CI 1.4 –2.4; P ⬍ 0.0001). increase in odds of change in therapy (OR 1.12 per 1 U in-
Patients were also more likely to have a change made in crease; 95% CI 1.05–1.19; P ⬍ 0.001). An increase in
outpatient therapy when they were receiving insulin ther- mean blood glucose by 10 mg/dl during admission was
apy before admission (OR 1.5; relative to no preadmission associated with a 7% increase in odds of change in therapy
2022 Griffith et al. Clinical Inertia of DM Discharge Planning J Clin Endocrinol Metab, June 2012, 97(6):2019 –2026

TABLE 2. Admission characteristics according to change in outpatient diabetes therapy at discharge

No change Outpatient DM therapy
Variable Na (n ⴝ 1571) changed (n ⴝ 454)
Age (yr) 2025 60 (55, 68) 59 (53, 68)
Site 2025
Huntington 354 108
Louisville 472 113
Tennessee Valley Healthcare System 745 233
Race 1973
White 1243 333
African-American 271 108
Other 18 0
Previous admission to VA within 90 d 2025 414 (26%) 97 (21%)
Length of stay (d) 2025 3 (1, 6) 4 (2, 7)
Most recent HgbA1c (%) 2025 9.5 (8.7, 10.8) 10 (8.9, 11.8)
Days from most recent HgbA1c to admission 2025 91 (41, 134) 78 (18, 128)
First inpatient creatinine (mg/dl) 1770 1.1 (0.9, 1.5) 1.1 (0.9, 1.5)
Maximal inpatient creatinine (mg/dl) 1773 1.2 (0.9, 1.7) 1.1 (0.9, 1.6)
Last inpatient creatinine (mg/dl) 1772 1.1 (0.9, 1.4) 1 (0.8, 1.3)
First inpatient glucose (mg/dl) 1988 195 (141, 262) 238 (166, 338)
Mean inpatient glucose (mg/dl) 1997 186 (154, 228) 214 (171, 256)
Last inpatient glucose (mg/dl) 1995 175 (126, 232) 182 (132, 243)
Inpatient hypoglycemia (⬍50 mg/dl) 1997 183 (12%) 90 (20%)
Intensive care unit admission 2025 272 (17%) 83 (18%)
Dialysis during admission 2025 20 (1%) 3 (1%)
Chronic kidney disease 2025 123 (8%) 37 (8%)
Inpatient endocrinology consult 2025 53 (3%) 67 (15%)
Preadmission insulin prescription 2025 917 (58%) 308 (68%)
Any inpatient insulin administration 2025 1,440 (92%) 434 (96%)
Basal insulin administration 2025 1,118 (71%) 392 (86%)
Hypoglycemic agent administered inpatient, other than insulin 2025 972 (62%) 282 (62%)
Admission service 2025
Medical 1,021 (65%) 364 (80%)
Surgical 269 (17%) 23 (5%)
Other 122 (8%) 32 (7%)
Mixed 159 (10%) 35 (8%)
DM, Diabetes mellitus.
a
Number of nonmissing values. Continuous variables are presented as medians with IQR (25th percentile, 75th percentile). Categorical variables
are presented as counts with percentages.

(OR 1.07 per 10 mg/dl increase; 95% CI 1.05–1.10; P ⬍
0.0001) and with occurrence of inpatient hypoglycemia
(blood glucose ⬍ 50 mg/dl; OR 1.82, relative to blood
glucose of 50 mg/dl or greater; 95% CI 1.32–2.51; P ⬍
0.001). Also, having inpatient basal insulin therapy in-
creased the odds of change in therapy by 71% (OR 1.71;
relative to no basal insulin therapy, 95% CI 1.25 ⫽
5–2.35; P ⬍ 0.01). The adjusted model showed no evi-
dence of association of change in therapy for Black race
(OR 1.15 relative to white race; 95% CI 0.86 –1.55; P ⫽
0.343), and no evidence that previous admission within
90 d was associated with change in therapy (OR 0.79;
relative to no previous admission, 95% CI 0.60 –1.04; P ⫽
0.098). Those patients admitted to nonmedical admission
services were less likely to have changes to therapy. Inpa-
tient endocrinology consultation showed strong associa-
tion with change in therapy (OR 4.27; relative to no en-
docrinology consultation, 95% CI 2.78 – 6.56; P ⬍
0.0001). The percentage of patients with no change in
diabetes therapy did not substantially vary over the time
period of the study (Fig. 1).
Among the 454 admissions marked by some change in
outpatient diabetes medications, 396 (87%) had at least
one medication change that included addition of a medi-
cation, increase in an existing medication dose, or a
change in type of insulin therapy. The most frequent
change was initiation of a new insulin agent, which oc-
curred among 202 of the 454 admissions with changes
made (44%). A new noninsulin agent was added in 78
cases (17%); 55 (12%) had a change from one type of
insulin to another, and 154 (34%) had dose of a previous
medication increased.
Of 2025 admissions, 647 (32%) were followed up by
an HgbA1c within 60 d of discharge. Forty-six percent of
all admissions (935 of 2025) had a follow-up appoint-
ment with primary care or a diabetes/endocrinology
clinic scheduled within 30 d of discharge. Of the ad-
missions with a change made to diabetes therapy, 145 of
J Clin Endocrinol Metab, June 2012, 97(6):2019 –2026 jcem.endojournals.org 2023

TABLE 3. Univariate analysis of associations with change in medications

Variable OR 95% CI P value
Age (per 1 yr) 0.99 0.99 –1.0 0.059
Previous admission to VA within 90 days 0.76 0.59 – 0.98 0.034
Adjusted length of stay (per 5 days) 0.99 0.97–1.02 0.64
Most recent HbgA1c (per 1 U) 1.19 1.11–1.26 ⬍0.0001
Maximal inpatient creatinine (per 0.1 mg/dl) 1.0 0.99 –1.01 0.4
Last inpatient creatinine (per 0.1 mg/dl) 0.99 0.98 –1.01 0.48
Mean inpatient glucose (per 10 mg/dl) 1.06 1.04 –1.08 ⬍0.0001
Last inpatient glucose (per 10 mg/dl) 1.01 0.99 –1.03 0.074
Days from start of study period to admission (per 1 yr) 1.04 0.99 –1.1 0.137
Race (referent, white)
African-American 1.5 1.13–1.96 0.005
Other 0.0002 0.0001– 0.0004 ⬍0.0001
Site (referent, Tennessee Valley)
Huntington 0.98 0.75–1.26 0.85
Louisville 0.77 0.59 – 0.99 0.045
Inpatient hypoglycemia (⬍50 mg/dl) 1.84 1.41–2.40 ⬍0.0001
Basal insulin administration 2.6 1.9 –3.4 ⬍0.0001
Inpatient endocrinology consultation obtained 4.96 3.39 –7.25 ⬍0.0001
Admission service (referent, medical)
Surgical 0.24 0.16 – 0.37 ⬍0.0001
Other (psychiatric, rehabilitation) 0.74 0.48 –1.13 0.157
Mixed 0.62 0.42– 0.92 0.017
Major diagnostic category (referent, endocrine/nutritional/ metabolic disorders)
Nervous system 0.11 0.05– 0.23 ⬍0.0001
Eye disorders 0.14 0.06 – 0.31 ⬍0.0001
Ear, nose, mouth, and throat 0.055 0.02– 0.13 ⬍0.0001
Respiratory system 0.11 0.05– 0.23 ⬍0.0001
Circulatory system 0.14 0.07– 0.29 ⬍0.0001
Digestive system 0.13 0.05– 0.3 ⬍0.0001
Hepatobiliary system and pancreas 0.05 0.02– 0.16 ⬍0.0001
Musculoskeletal/connective tissue 0.09 0.04 – 0.2 ⬍0.0001
Skin, sc tissue and breast 0.16 0.07– 0.39 ⬍0.0001
Kidney and urinary tract 0.1 0.04 – 0.28 ⬍0.0001
Mental diseases and disorders 0.14 0.06 – 0.31 ⬍0.0001
Alcohol/drug use and alcohol/drug-induced organic mental disorders 0.14 0.06 – 0.33 ⬍0.0001
Other 0.12 0.05– 0.27 ⬍0.0001
Preadmission insulin prescription 1.5 1.2–1.9 ⬍0.001
Any inpatient insulin administration 1.97 1.23–3.18 0.005
Chronic kidney disease 1.04 0.7–1.5 0.82
Intensive care unit admission 1.07 0.82–1.39 0.63

454 (32%) had an HgbA1c collected within 60 d of
discharge. Median postdischarge HgbA1c for this
group was 8.7% (IQR 7.6 –10.7%). Two hundred fifty
of 454 (55%) had a follow-up appointment with pri-
mary care or endocrinology scheduled within 30 d of
discharge. After the admissions without a change in
diabetes therapy, 502 of 1571 (32%) had an HgbA1c
within 60 d of discharge, with median postdischarge
HgbA1c 7.9 (IQR 7.0 –9.1%). Six hundred eighty-five
of 1571 (44%) were scheduled for a follow-up appoint-
ment with primary care or endocrinology within 30 d of
discharge. Among those patients not scheduled for fol-
low-up within 30 d and without a change in therapy, the
mean inpatient glucose was 197 mg/dl (SD 58 mg/dl).
However, 656 of the cohort of 2025 admissions (32%),
all with uncontrolled HgbA1c at admission, demon-
strated aggregate clinical inertia with no change in di-
abetes therapy, no documentation of HgbA1c within
60 d of discharge, and no follow-up appointment sched-
uled within 30 d of discharge.
Conclusions
In this multicenter, retrospective study of patients with
poorly controlled diabetes and at least one hospitalization,
we found that less than a quarter received a change in out-
patient diabetes therapy upon discharge. These data indicate
significant clinical inertia for inpatient clinical teams to avoid
making any changes in a patient’s diabetes regimen at dis-
charge, even when it is clear from HgbA1c data that the
preexisting outpatient regimen was ineffective. In addition, a
third of these patient admissions did not receive a change in
outpatient diabetes medications, an outpatient follow-up ap-
pointment within 30 d, or follow-up HgbA1c within 60 d
after discharge. The rate of clinical inertia did not substan-
2024 Griffith et al. Clinical Inertia of DM Discharge Planning J Clin Endocrinol Metab, June 2012, 97(6):2019 –2026

TABLE 4. Multivariable analysis of associations with change in medications. CI ⫽ confidence interval

Variable OR 95% CI P value
Age (per 1 yr) 1.00 0.98 –1.01 0.92
Previous admission to VA within 90 d 0.79 0.60 –1.04 0.098
Adjusted length of stay (per 5 d) 0.98 0.94 –1.01 0.134
Most recent HgbA1c (per 1 U) 1.12 1.05–1.19 ⬍0.001
Maximal inpatient creatinine (per 0.1 mg/dl) 0.99 0.97–1.01 0.193
Last inpatient creatinine (per 0.1 mg/dl) 1.01 0.99 –1.03 0.407
Mean inpatient glucose (per 10 mg/dl) 1.07 1.05–1.10 ⬍0.0001
Last inpatient glucose (per 10 mg/dl) 0.98 0.97–1.00 0.057
Days from start of study period to admission (per 1 yr) 0.95 0.90 –1.01 0.128
Race (referent, white)
African-American 1.15 0.86 –1.55 0.343
Other 0.0004 0.0 – 0.04 ⬍0.001
Site (referent, Tennessee Valley)
Huntington 0.83 0.62–1.13 0.237
Louisville 0.79 0.60 –1.04 0.092
Inpatient hypoglycemia (⬍50 mg/dl) 1.82 1.32–2.52 ⬍0.001
Basal insulin administration 1.71 1.25–2.35 ⬍0.001
Inpatient endocrinology consultation obtained 4.27 2.78 – 6.56 ⬍0.0001
Admission service (referent, medical)
Surgical 0.24 0.16 – 0.37 ⬍0.0001
Other (psychiatric, rehabilitation) 0.36 0.14 – 0.91 0.031
Mixed 0.53 0.34 – 0.82 0.005
Major diagnostic category (referent, endocrine/nutritional/ metabolic disorders)
Nervous system 0.14 0.06 – 0.33 ⬍0.0001
Eye disorders 0.19 0.08 – 0.48 ⬍0.001
Ear, nose, mouth, and throat 0.10 0.04 – 0.27 ⬍0.0001
Respiratory system 0.15 0.06 – 0.34 ⬍0.0001
Circulatory system 0.21 0.10 – 0.45 ⬍0.0001
Digestive system 0.26 0.10 – 0.66 0.005
Hepatobiliary system and pancreas 0.07 0.02– 0.23 ⬍0.0001
Musculoskeletal/connective tissue 0.19 0.08 – 0.44 ⬍0.0001
Skin, subcutaneous tissue and breast 0.23 0.09 – 0.62 0.003
Kidney and urinary tract 0.19 0.06 – 0.54 0.002
Mental diseases and disorders 0.27 0.08 – 0.91 0.035
Alcohol/drug use and alcohol/drug-induced organic mental disorders 0.30 0.10 – 0.92 0.036
Other 0.19 0.07– 0.49 ⬍0.001

tially change over the study period. Taken together,
these data present a clear opportunity to improve the
management of poorly controlled diabetes at hospital
discharge and with outpatient follow-up.
We identified a number of clinical surveillance vari-
ables that were associated with the alteration of outpatient
FIG. 1. Percentage of patient admissions with uncontrolled diabetes
per calendar year discharged with no change in diabetes medications,
no follow-up appointment within 30 d, and no follow-up HgbA1c
within 60 d by year of study.
diabetes therapy at hospital discharge, including higher
HgbA1c values, higher mean inpatient glucose, the occur-
rence of hypoglycemia during the hospital admission, and
use of basal insulin during the hospital admission. These
factors represent clinically reasonable indications for a
change in therapy, and some were consistent with findings
of other studies of treatment intensification. In the Trans-
lating Research into Action for Diabetes study, 48% of
1093 patients with suboptimal control of type 2 diabetes
(HgbA1c ⬎ 7.2%) experienced intensification of therapy
with an addition of an oral medication or insulin. Younger
age and higher baseline HgbA1c predicted treatment in-
tensification in the authors’ adjusted multivariable anal-
ysis (20). In another study of VA patients, outpatient in-
tensification of diabetes therapy was more likely if the
most recent HgbA1c was greater than 8.0% and with
higher serum glucose, more recent HgbA1c, insulin use,
and having glucose self-monitoring supplies (9).
We found that patients admitted for acute medical con-
ditions other than endocrine, nutritional, or metabolic dis-
orders were less likely to have change to diabetes therapy
upon discharge. The sole exclusion among other catego-
ries was psychiatric admissions, although these were rel-
J Clin Endocrinol Metab, June 2012, 97(6):2019 –2026
atively few in number. This indicates diffuse inattention to
diabetes during inpatient admissions when it is not the
primary focus, even when there is evidence of chronically
poor control. Patients admitted to surgical, psychiatric, or
rehabilitation services were less likely to have a change in
subsequent outpatient therapy compared with patients
admitted to medical services. Hyperglycemia in surgical
patients is associated with increased risk of surgical site
infections as well as increases in mortality, with strong
data for specific populations such as postcardiac surgery
patients (21, 22). Our data suggest that providers on non-
medical services may not feel as comfortable as medical
services with making a change to diabetes medications
after patient discharge. Some providers may believe it is
inappropriate to change an outpatient regimen managed
by another physician. This population’s high rate of in-
sulin use preadmission may have contributed to reluctance
to alter a regimen if an inpatient provider lacks extensive
experience with insulin titration. However, in such cases,
short-term outpatient follow-up should be expected; less
than half of patients without a therapy change had a fol-
low-up appointment scheduled within 30 d.
Prior studies have found racial disparity for treatment
intensification among patients with diabetes. Some prior
studies found that white patients were more likely to re-
ceive diabetic treatment intensification, whereas others
have shown that nonwhite patients experienced greater
improvements in HgbA1c measurements after treatment
intensification (20, 23, 24). Within the VA, racial dispar-
ities appear to be minimal in recent years (25). Consistent
with this, we did not see a significant difference in changes
made for African-American patients compared with white
patients. Nonwhite, non-African-American patients ap-
peared to be less likely to experience change in therapy, but
this was a very small number of patients, so we were cau-
tious about drawing conclusions for that group.
Poor patient adherence has been shown to be a factor
in clinical inertia, with providers less likely to make
changes over 1 yr of clinical outpatient follow-up after an
elevated HgbA1c for the least adherent patients (7). This
study design did not allow for complete assessment of
patient adherence. We used patients’ preadmission pre-
scriptions filled, as opposed to issued, for the assessment
of their baseline medication regimen to partially account
for adherence.
Discharge planning recommendations within the 2011
American Diabetes Association standards of care include
having a follow-up within 1 month with a primary or
diabetes care provider (such as primary care provider, en-
docrinologist, or diabetes educator) for all patients who
are hyperglycemic in the hospital (13). Early postdischarge
telephone follow-up with diabetes nurse specialists has
jcem.endojournals.org 2025
been shown to improve HgbA1c at 24 wk after discharge,
even in patients with suboptimal glycemic control at dis-
charge (26). Early follow-up attention to diabetes should
be considered in discharge planning for all patients who
presented with poor control or had significant hypergly-
cemia as inpatients.
Inpatient consultation from a medical or endocrine ser-
vice may also help address follow-up planning more ef-
fectively if provider comfort with outpatient diabetes
management is affecting discharge planning. In this co-
hort, only 6% of all admissions had endocrine inpatient
consultation, but there was a strong adjusted association
with changes made to outpatient therapy for those pa-
tients receiving consultation.
This study had a number of limitations, including those
inherent to a retrospective cohort study design. Women
were excluded from the cohort due to having a low num-
ber of eligible female cases; thus, it is uncertain whether
similar findings would occur in female patients. However,
a large difference in diabetes discharge planning based on
patient sex seems unlikely. Other studies of clinical inertia
in diabetes have not shown a linkage to patient sex (5–7).
We were unable to assess features such as the clinical his-
tory collected from patients by providers at admission
regarding recent diabetes parameters such as hypogly-
cemia. We also could not evaluate discharge counseling
regarding medication or dietary adherence. Any fol-
low-up appointments made with non-VA providers
could not be assessed, but we attempted to mitigate this
by including only patients who were prescribed diabetes
medications within the VA system. The study was not
designed to explain provider behavior, only to describe
changes made to therapy.
Overall, this study suggests many areas in which atten-
tion to outpatient diabetes care can be improved by ad-
dressing it during hospital admission and discharge plan-
ning. We noted a low percentage of patients with primary
care or diabetes appointments made after discharge in ad-
dition to a low rate of change to therapy. Adjustment of the
outpatient diabetes regimen may be indicated by poor pre-
admission control parameters. Consultation from an in-
patient endocrinology or diabetes service may facilitate
inpatient diabetes care as well as discharge planning. Ad-
ditionally, attention to appropriate short-term follow-up
may improve the transition to outpatient care and con-
tinued improvement in care of this chronic condition. A
promising area of future study will be implementation and
evaluation of interventions to improve diabetes discharge
planning and therapy adjustment in hospital patients with
uncontrolled diabetes, to determine if such interventions
can improve glycemic control.
2026 Griffith et al. Clinical Inertia of DM Discharge Planning
Acknowledgments
We thank Arijit Basu and Vincent Messina (Geriatric Research,
Education, and Clinical Care Center, Veterans Affairs Tennessee
Valley Health System), for research support activities. The fol-
lowing authors contributed to this manuscript: M.L.G. re-
searched the data, contributed to the study design and analysis,
and wrote the manuscript. J.B.B. contributed to the study design
and analysis, contributed to the discussion, and reviewed/edited
the manuscript. S.K.E. performed the statistical analysis, wrote
portions of the manuscript, and reviewed/edited the manuscript.
M.E.M. contributed to the study design and directed the anal-
ysis, contributed to the discussion, and reviewed/edited the
manuscript.
Address all correspondence and requests for reprints to: Mi-
chael E. Matheny, M.D., M.S., M.P.H., Assistant Professor of
Medicine, Biostatistics, and Bioinformatics, Veterans Affairs
Tennessee Valley Healthcare System, A-413 VA/GRECC, 1310
24th Avenue South, Nashville, Tennessee 37212. E-mail:
michael@matheny.info.
M.E.M. was supported by the Veterans Health Administra-
tion Health Services Research and Development Career Devel-
opment Award CDA-08-020. This material is the result of work
supported with resource and the use of facilities within the De-
partment of Veterans Affairs.
Disclosure Summary: The authors have no relevant conflicts
of interests to disclose.
References
1. Miller DR, Safford MM, Pogach LM 2004 Who has diabetes? Best
estimates of diabetes prevalence in the Department of Veterans Af-
fairs based on computerized patient data. Diabetes Care 27(Suppl
2):B10 –B21
2. Holman RR, Paul SK, Bethel MA, Matthews DR, Neil HA 2008
10-year follow-up of intensive glucose control in type 2 diabetes.
N Engl J Med 359:1577–1589
3. Falanga V 2005 Wound healing and its impairment in the diabetic
foot. Lancet 366:1736 –1743
4. Umpierrez GE, Hellman R, Korytkowski MT, Kosiborod M, May-
nard GA, Montori VM, Seley JJ, Van den Berghe G 2012 Manage-
ment of hyperglycemia in hospitalized patients in non-critical care
setting: an endocrine society clinical practice guideline. J Clin En-
docrinol Metab 97:16 –38
5. Phillips LS, Twombly JG 2008 It’s time to overcome clinical inertia.
Ann Intern Med 148:783–785
6. Ziemer DC, Miller CD, Rhee MK, Doyle JP, Watkins Jr C, Cook CB,
Gallina DL, El-Kebbi IM, Barnes CS, Dunbar VG, Branch Jr WT,
Phillips LS 2005 Clinical inertia contributes to poor diabetes control
in a primary care setting. Diabetes Educ 31:564 –571
7. Grant R, Adams AS, Trinacty CM, Zhang F, Kleinman K, Soumerai
SB, Meigs JB, Ross-Degnan D 2007 Relationship between patient
medication adherence and subsequent clinical inertia in type 2 dia-
betes glycemic management. Diabetes Care 30:807– 812
8. Grant RW, Cagliero E, Dubey AK, Gildesgame C, Chueh HC, Barry
J Clin Endocrinol Metab, June 2012, 97(6):2019 –2026
MJ, Singer DE, Nathan DM, Meigs JB 2004 Clinical inertia in the
management of type 2 diabetes metabolic risk factors. Diabet Med
21:150 –155
9. Berlowitz DR, Ash AS, Glickman M, Friedman RH, Pogach LM,
Nelson AL, Wong AT 2005 Developing a quality measure for clin-
ical inertia in diabetes care. Health Serv Res 40(6 Pt 1):1836 –1853
10. Matheny ME, Shubina M, Kimmel ZM, Pendergrass ML, Turchin
A 2008 Treatment intensification and blood glucose control among
hospitalized diabetic patients. J Gen Intern Med 23:184 –189
11. Knecht LA, Gauthier SM, Castro JC, Schmidt RE, Whitaker MD,
Zimmerman RS, Mishark KJ, Cook CB 2006 Diabetes care in the
hospital: is there clinical inertia? J Hosp Med 1:151–160
12. Cook CB, Castro JC, Schmidt RE, Gauthier SM, Whitaker MD,
Roust LR, Argueta R, Hull BP, Zimmerman RS 2007 Diabetes care
in hospitalized noncritically ill patients: more evidence for clinical
inertia and negative therapeutic momentum. J Hosp Med 2:203–
211
13. Cook CB, Seifert KM, Hull BP, Hovan MJ, Charles JC, Miller-Cage
V, Boyle ME, Harris JK, Magallanez JM, Littman SD 2009 Inpatient
to outpatient transfer of diabetes care: planning for an effective
hospital discharge. Endocr Pract 15:263–269
14. American Diabetes Association 2011 Standards of medical care in
diabetes—2011. Diabetes Care 34(Suppl 1):S11–S61
15. Williams RL 2000 A note on robust variance estimation for cluster-
correlated data. Biometrics 56:645– 646
16. Huber P 1967 The behavior of maximum likelihood estimates under
nonstandard conditions. Proc Fifth Berkeley Symposium Math Stat
1:221–233
17. White H 1982 Maximum likelihood estimation of misspecified
models. Econometrica 50:1–25
18. Little RJA, Rubin DB 2002Statistical analysis with missing data.
2nd ed. New York: John Wiley
19. Rubin DB, Schenker N 1991 Multiple imputation in health-care
databases: an overview and some applications. Stat Med 10:585–
598
20. McEwen LN, Bilik D, Johnson SL, Halter JB, Karter AJ, Mangione
CM, Subramanian U, Waitzfelder B, Crosson JC, Herman WH
2009 Predictors and impact of intensification of antihyperglycemic
therapy in type 2 diabetes: translating research into action for dia-
betes (TRIAD). Diabetes Care 32:971–976
21. Ata A, Lee J, Bestle SL, Desemone J, Stain SC 2010 Postoperative
hyperglycemia and surgical site infection in general surgery patients.
Arch Surg 145:858 – 864
22. Furnary AP, Wu Y 2006 Clinical effects of hyperglycemia in the
cardiac surgery population: the Portland Diabetic Project. Endocr
Pract 12(Suppl 3):22–26
23. Rodondi N, Peng T, Karter AJ, Bauer DC, Vittinghoff E, Tang S,
Pettitt D, Kerr EA, Selby JV 2006 Therapy modifications in response
to poorly controlled hypertension, dyslipidemia, and diabetes mel-
litus. Ann Intern Med 144:475– 484
24. Chou AF, Brown AF, Jensen RE, Shih S, Pawlson G, Scholle SH 2007
Gender and racial disparities in the management of diabetes mellitus
among Medicare patients. Womens Health Issues 17:150 –161
25. Trivedi AN, Zaslavsky AM, Schneider EC, Ayanian JZ 2005 Trends
in the quality of care and racial disparities in Medicare managed
care. N Engl J Med 353:692–700
26. Wong FK, Mok MP, Chan T, Tsang MW 2005 Nurse follow-up of
patients with diabetes: randomized controlled trial. J Adv Nurs 50:
391– 402