In recent decades, the prevalence of diabetes has been increasing in many countries, including Singapore.13  Consequently, a large number of hospital admissions are related either directly or indirectly to diabetes or to one of its many complications. Approximately 30% of admissions to our tertiary hospital are diabetes related (based on internal audit data).

Insulin therapy is required for the management of hyperglycemia in patients with or without diabetes. The main danger with insulin use is life-threatening hypoglycemia. Insulin is one of the five “high-alert” medications identified by the Joint Commission on Accreditation of Healthcare Organizations to have a significant risk of injurious consequences in the event of a medication error.4 

Insulin therapy has grown more complex with the advent of insulin analogs and new insulin regimens, and the presence of knowledge gaps among health care professionals (HCPs) has been one of the important factors resulting in errors in insulin use.5  This is compounded by similar-sounding trade or generic insulin names (e.g., “lente” [an intermediate-acting human insulin formulation] and “Lantus” [the trade name for the long-acting insulin analog glargine])6  and by systems in which different HCPs are responsible for prescribing and administering insulin.7 

Physicians, nurses, and pharmacists all play important roles in inpatient insulin use. Studies in Europe, America, and South Asia have shown significant deficits in insulin-related knowledge among health care workers.810  Nurses have been found to have inadequate knowledge to teach patients basic insulin administration skills,8,9  and physicians have been found to give suboptimal diabetes care (based on American Diabetes Association [ADA] guidelines).10  However, to our knowledge, there are no published studies on insulin-related knowledge from Singapore.

In this study, we used an anonymous survey to assess insulin-related knowledge among HCPs at a tertiary hospital in Singapore, including physicians and nurses working in internal medicine, as well as pharmacists. In this way, we aimed to identify differences in insulin-related knowledge among the different professions and to look for specific knowledge gaps in insulin-related knowledge. In so doing, we hoped to be able to make recommendations for system changes and education programs to improve patient care and safety by decreasing the risk of insulin-related medical errors.

A questionnaire was prepared based on a similar form used in a study by Derr et al.11  The questions were adapted to the most recent ADA clinical guidelines available at that time12  and to the types of insulin preparations available within the local health service. Drafts of the questionnaire were reviewed collectively by the hospital's Endocrinology Department to ensure relevance, accuracy, and ease of understanding.

The final questionnaire comprised 14 knowledge-based questions (Table 1), with 7 focused on insulin characteristics and prescription and 7 focused on insulin preparation and administration. These 14 multiple-choice questions had four options with a single best answer for each. Participants were also asked for information about their profession, the number of years they had been in clinical practice, and their self-described level of comfort in managing patients with diabetes.

Table 1.

List of Knowledge-Based Questions

List of Knowledge-Based Questions
List of Knowledge-Based Questions

The questionnaire was administered to physicians, nurses, and pharmacists at a large tertiary hospital in Singapore. The hospital staff included 323 physicians from medical departments, 603 registered nurses from medical wards, and 47 inpatient pharmacists.

The questionnaire was administered to physicians and pharmacists during small group meetings such as journal clubs or faculty or departmental meetings and to nurses during their shift changes from March 2012 to April 2012. At each session, one of the investigators would be present to distribute and brief the participants on the questionnaire and collect the completed questionnaires.

Participation in the study was voluntary and anonymous. All questionnaires were completed in one sitting. No time limit was given, although participants were not permitted to communicate with each other or to refer to any reference material. Care was taken to ensure no duplicate submissions.

Completed questionnaires were scored manually and double-entered into a database. Any discrepancies were counterchecked against the original survey form. Questions that were left blank or had more than one answer chosen were considered to be incorrect.

Statistical analyses were performed using Excel 2007 and Stata 10.0 (Microsoft Singapore Pte. Ltd., Singapore). Medians and interquartile ranges (IQRs) are used throughout to present the scores for the knowledge-based questions because the distribution of the scores was nonparametric. Hypothesis testing was carried out using the Mann-Whitney U test (2 groups) or the Kruskal-Wallis test (≥ 2 groups). Correction of P values for multiple comparisons using the Bonferroni method was carried out where appropriate.

Ethics approval was obtained from the Domain-Specific Review Boards (the National Healthcare Group).

Study population

A total of 417 questionnaires were distributed to 160 physicians from medical departments, 226 registered nurses from medical wards, and 31 inpatient pharmacists. Whether doctors, nurses, and pharmacists were offered the survey was dependent on their presence at department meetings or nursing shift changes. Of the 417 questionnaires given out, 375 were completed and analyzed (138 from physicians, 209 from nurses, and 28 from pharmacists), giving an overall response rate of 89.9% (86.3% for physicians, 92.5% for nurses, and 90.3% for pharmacists).

Comfort level

The self-reported comfort level of survey respondents in managing diabetes showed that 84.7% of physicians, 95.7% of nurses, and 82.1% of pharmacists reported being “very” or “somewhat” comfortable in managing diabetes. Among the physicians and nurses, the percentages reporting that they were “very” or “somewhat” comfortable in managing diabetes were similar across experience levels.

Table 2.

Median Scores for Knowledge-Based Questions Categorized by Profession and Years in Clinical Practice

Median Scores for Knowledge-Based Questions Categorized by Profession and Years in Clinical Practice
Median Scores for Knowledge-Based Questions Categorized by Profession and Years in Clinical Practice
Figure 1.

Graphs of median scores for knowledge-based questions by question category and profession (horizontal marker indicates median, boxes indicate interquartile range, and vertical bars indicate range).

Figure 1.

Graphs of median scores for knowledge-based questions by question category and profession (horizontal marker indicates median, boxes indicate interquartile range, and vertical bars indicate range).

Close modal

Knowledge scores according to profession, years of practice, and comfort level

Table 2 and Figure 1 show the median scores and IQRs for the knowledge-based questions, categorized by profession.

Total knowledge scores were found to vary with profession. Pharmacists scored the highest median score of 12 (IQR 10.5–12); physicians scored a median score of 8 (IQR 6–10); and nurses scored a median score of 7 (IQR 6–8). All pairwise comparisons between physicians, nurses, and pharmacists were statistically significant, yielding corrected P values of < 0.001.

Scores for questions on insulin characteristics and prescription were found to be similar among physicians and pharmacists (both with median scores of 5, IQR 4–6) but lower in nurses (median score of 2, IQR 1–3). Pairwise comparisons between physicians and nurses as well as between pharmacists and nurses yielded corrected P values of < 0.001.

Scores for questions on insulin preparation and administration were lowest among physicians (median score 4, IQR 3–5) when compared to nurses (median score 5, IQR 4–6) and pharmacists (median score 6, IQR 6–7). Again, all pairwise comparisons between physicians, nurses, and pharmacists yielded corrected P values of < 0.001.

When scores were analyzed based on respondents' number of years in clinical practice, it was found that physicians with ≥ 11 years of clinical practice had lower median total scores (median score 7, IQR 4–10) compared to physicians with ≤ 5 and 6–10 years of clinical practice, who had similar total scores (median score 9, IQR 7–11). This difference was statistically significant when the most experienced physicians were compared separately with the other two groups of physicians (both corrected P values < 0.05). Among nurses, however, total scores were not associated with the number of years of practice (P > 0.5).

Total scores were not significantly associated with the respondents' self-reported comfort level in managing diabetes, either in the overall population or for physicians and nurses analyzed separately (P = 0.35 overall, P = 0.93 for physicians, and P = 0.73 for nurses). Pharmacists were not analyzed by comfort level because of the small number of pharmacist respondents.

Table 3.

Percentage Correct for Questions About Insulin Characteristics and Prescription

Percentage Correct for Questions About Insulin Characteristics and Prescription
Percentage Correct for Questions About Insulin Characteristics and Prescription

Table 3 shows the percentage of correct answers for the individual questions on insulin characteristics and prescription, and Table 4 shows the percentage of correct answers for the individual questions on insulin preparation and administration.

Discussion

Insulin is one of the four drugs that cause the most adverse events during hospital stays for seniors in the United States, along with warfarin, oral antiplatelets, and oral hypoglycemic agents.13  Despite this, studies have continued to show deficiencies in health care professionals' knowledge of insulin use.6,8,10  This study aimed to quantify the extent of this problem among health care workers and attempted to show the differences in understanding of and knowledge about insulin use among physicians, nurses, and pharmacists.

Differences in knowledge between professions

We found differences in insulin-related knowledge among the different professions. The overall median knowledge score (out of 14) was 12 for pharmacists, 8 for physicians, and 7 for nurses. Pharmacists scored the highest in both categories of insulin prescription and administration, likely because of their rigorous theoretical and practical training in medication use. Physicians scored better than nurses for questions related to insulin characteristics and prescription, whereas nurses scored better on questions related to insulin preparation and administration.

In most hospitals, physicians are responsible for prescribing medications, pharmacists for dispensing them, and nurses for administering them. The differences we observed in this study are likely the result of differences in job scope between physicians and nurses. As an example, only 44.9% of physicians knew that 1 ml of standard-concentration insulin contains 100 units of insulin. This could lead to life-threatening consequences such as hypoglycemic coma if physicians with this knowledge deficit were to prepare and administer insulin without assistance in emergencies such as severe hyperkalemia.

Poor knowledge of newer insulin analogs

It was observed that physicians scored the most poorly with regard to questions involving the relatively new insulin analog detemir. In addition, the more years of clinical practice physicians had, the less likely they were to answer these questions correctly.

We postulate that this is because detemir is one of the newer insulin analogs introduced to our institution, and as a result, senior physicians may not have encountered it in their years of formal training. It is important for all physicians to keep up-to-date with developments in insulin analogs because, although they may not initiate the prescription of these analogs, they may be required to continue or adjust doses of these analogs as part of patients' preexisting medication regimen.

Table 4.

Percentage Correct for Questions About Insulin Preparation and Administration

Percentage Correct for Questions About Insulin Preparation and Administration
Percentage Correct for Questions About Insulin Preparation and Administration

Lack of correlation between comfort level and knowledge

Interestingly, among both physicians and nurses, we found no statistical correlation between their comfort level in managing diabetes and their insulin survey scores. Although insulin is well known and commonly used in the management of hospitalized patients and > 80% of respondents reported being comfortable in managing diabetes, familiarity does not equate to better knowledge.

Recommendations for improvement

Apart from education to improve the knowledge of all three professions, an important safeguard to prevent medication errors would be to ensure that physicians, nurses, and pharmacists communicate effectively as a team. Harnessing the knowledge strengths of each profession would improve the possibility of detecting prescription, transcription, or dispensing errors before patient harm occurs. Increased redundancy should be built into the system by giving pharmacists a role of ascertaining safe insulin use and highlighting any concerns to physicians and nurses.

Even with these safeguards, we recommend that the onus should still lie with prescribers to ensure that clear instructions are passed to the rest of the team. The use of electronic systems to prescribe inpatient medications can help facilitate this process by providing prewritten order sets and decreasing transcription errors.14 

Continuing medical education and the provision of timely updates of information about new insulin products or their variants introduced into a hospital system are essential. A systematic approach to assess insulin-related knowledge and disseminate insulin-related information should be undertaken by a multidisciplinary team, with participation from all health care professions involved in caring for patients with diabetes. This team should also help to put in place measures to ensure that the interpretation of insulin orders is consistent among all health care workers.

Limitations of the study

Several limitations in our study merit discussion. Our questionnaire has not been validated formally and was only reviewed by endocrinologists before administration without being pilot-tested among general medical physicians, nurses, and pharmacists. As a result, different interpretations of the questions could have affected the study results.

Nonresponse bias is also a possibility, particularly for physicians, because the questionnaire was administered during department meetings and journal clubs, where attendance is often not complete. However, the effect of this is not likely to be great given the nearly 90% response rate among those offered the questionnaire. Although logistical constraints prevented offering the questionnaire to all relevant physicians, nurses, and pharmacists in the hospital, we do not expect any major differences between those who were and were not offered the questionnaire.

Despite these limitations, we believe the results of the survey are reflective of the level of insulin-related knowledge within our institution.

Conclusion

This study has highlighted significant knowledge gaps about insulin use among HCPs. In addition to continuing medical education to address these knowledge deficits, more work should be done to facilitate communication among physicians, nurses, and pharmacists. Their interrelated roles and specific skill sets are essential for safe insulin use from the initial prescription to the eventual administration of insulin to patients at the bedside. As Sir William Osler wisely said, “Half of us are blind, few of us feel, and we are all deaf.”15  To move forward in terms of patient safety, we must recognize our inherent inadequacies.

The authors acknowledge Dr. Annabelle Rodriguez of Johns Hopkins Bayview Medical Center, who kindly allowed the adaptation of her questionnaire for this study.

1.
Wild
S
,
Roglic
G
,
Green
A
,
Sicree
R
,
King
H
:
Global prevalence of diabetes: estimates for 2000 and projections for 2030
.
Diabetes Care
27
:
1047
1053
,
2004
2.
Heng
BH
,
Sun
Y
,
Cheah
TS
,
Jong
M
: T
he Singapore National Healthcare Group diabetes registry: descriptive epidemiology of type 2 diabetes mellitus
.
Ann Acad Med Singapore
39
:
348
352
,
2010
3.
Singapore Ministry of Health
Epidemiology and Disease Control Division
:
National health survey 2010
.
Singapore
,
Ministry of Health
,
2010
. Available from http://www.singstat.gov.sg/Publications/publications_and_papers/health/ssnmar10-pg25-27.pdf.
Accessed 5 June 2012
4.
‘High-Alert’ medications and patient safety
.
Int J Qual Health Care
213
:
339
340
,
2001
5.
Hirsch
IB
:
Insulin analogues
.
N Engl J Med
352
:
174
183
,
2005
6.
Berkowitz
K
:
Lantus? Or Lente?
Am J Nurs
102
:
55
59
,
2002
7.
Page
MD
,
Stephenson
C
,
Pope
RM
,
Bodansky
HJ
:
Prescribing and dispensing of insulin: margins for error?
Diabet Med
9
:
938
941
,
1992
8.
El-Deraiwi
KM
,
Zuraikat
N
:
Registered nurses actual and perceived knowledge of diabetes mellitus
.
J Nurses Staff Dev
17
:
5
11
,
2001
9.
Baxley
S
,
Brown
S
,
Pokorny
M
,
Swanson
M
:
Perceived competence and actual knowledge of diabetes mellitus among nurses
.
J Nurses Staff Dev
13
:
93
98
,
1997
10.
Bernard
A
,
Anderson
L
,
Cook
C
,
Philips
L
:
What do internal medicine residents need to enhance their diabetes care
.
Diabetes Care
22
:
661
666
,
1999
11.
Derr
RL
,
Sivanandy
MS
,
Bronich-Hall
L
,
Rodriguez
A
:
Insulin-related knowledge among health care professionals in internal medicine
.
Diabetes Spectrum
20
:
177
185
,
2007
12.
American Diabetes Association
:
Standards of medical care in diabetes–2010
.
Diabetes Care
33
:
11
61
,
2010
13.
Budnitz
DS
,
Lovegrove
MC
,
Shehab
N
,
Richards
CL
:
Emergency hospitalizations for adverse drug events in older Americans
.
N Engl J Med
365
:
2002
2012
,
2011
14.
Shamliyan
TA
,
Duval
S
,
Du
J
,
Kane
RL
:
Review of the evidence of the impact of computerized physician order entry system (CPOE) on medication errors
.
Health Serv Res
43
:
32
53
,
2008
15.
Huth
EJ
,
Murrary
TJ
, Eds.
Medicine in Quotations
. 2nd ed.
Philadelphia, Pa.
,
American College of Physicians
,
2006