Exploration of diabetes knowledge among registered nurses working in an NHS Trust

The National Diabetes Inpatient Audit carried out in 2015 (Health and Social Care Information Centre [HSCIC], 2016) highlighted some improvements in diabetes care in the past 5 years, but identified a clear need for more. Satisfaction has not improved since the audit started in 2010, with some inpatients reporting staff had insufficient knowledge (HSCIC, 2016). This article focuses on a service evaluation of diabetes knowledge and training preferences among registered nurses (RNs) carried out as part of a master’s degree dissertation. 

Literature review
Various studies have investigated diabetes knowledge and training among nurses (e.g. Cytryn et al, 2009; Graue et al, 2010; Livingston and Dunning, 2010; King et al, 2012). Overall, they have shown suboptimal knowledge, the need for improvement, the benefits of diabetes training among nurses and the impact this can have on patient care. Modic et al (2013) found nurses were not confident or satisfactorily prepared to make decisions about diabetes management in hospital. Ahmed et al (2012) found diabetes management knowledge was significantly lacking among nurses in inpatient and outpatient settings. Strider and Phillips (2011) revealed deficient knowledge about hypoglycaemia and its management, which was affecting the quality and safety of care that at-risk patients received. Carney et al (2013) found gaps in nurses’ and student nurses’ knowledge of nutritional management of diabetes. 

There is a need to acknowledge nurses may lack insight into their level of diabetes knowledge and skills (Strider and Phillips, 2011). Gerard et al (2010) demonstrated mediocre levels of actual knowledge despite positive levels of perceived knowledge, with significant variations across treatments, the management of acute complications, long-term complications, diet and foot care. Ahmed et al (2012) stress the importance of investigating nurses’ actual knowledge to inform diabetes education development. They assert the need for awareness of knowledge plateaus and highlight the importance of reinforcing diabetes knowledge. Livingston and Dunning’s study (2010) reinforced the need for ongoing training and the use of up-to-date, reliable information sources.

Acquisition of knowledge does not automatically equate with better diabetes care (Holmes and Dyer, 2013). Education needs a strong link to practice to facilitate the transfer of evidence-based skills and knowledge into practice. Barriers such as “fear of change and fear of negative judgements […] competing demands […] lack of time […] workplace environment and structural and organisational barriers”, have been identified (Graue et al, 2010). 

Methodology
A questionnaire designed and piloted prior to this study was used to ascertain RNs’ diabetes knowledge and identify factors that might affect levels of knowledge. It was developed to enable specific characteristics pertinent to the Trust (management of hypoglycaemia and insulin management) to be investigated as part of a service evaluation. These were areas that personal experience, local incident reporting and results from the National Diabetes Inpatient Audit (HSCIC, 2014) highlighted as needing improvement. Participants were asked to rate statements relating to diabetes knowledge, confidence and experiences regarding diabetes care from “strongly disagree” to “strongly agree”.

Using SurveyMonkey’s sample size calculator, of the 742 RNs working with adults in the Trust, 86 RNs needed to participate in the study, allowing a 10% margin of error and 95% confidence interval. Questionnaires were distributed to 304 nurses to allow for a significant non-response rate. 

Data were analysed using descriptive statistics and SPSS software was used to perform non-parametric statistical tests. Nominal and ordinal level data were collected; therefore non-parametric statistical tests were applied (Bowling, 2014). Fisher’s exact test was used to determine whether there was a non-random relationship between two categorical variables and Kendall’s tau-b (t_b) correlation test (Fields, 2009) was used to determine the significance of participants’ answers (P<0.05 was considered statistically significant). 

Authorisation was sought through the NHS Trust’s Research and Development Department. The project was deemed to be a service evaluation, so Research NHS Ethics approval was not required. Participants were provided with a consent form, which was returned separately to the completed questionnaire to ensure the anonymity of responses. 

Results
Sixty-nine of the 304 questionnaires distributed were returned, a response rate of 22.7%. Sixty-eight participants gave valid answers for age and gender (see Figure 1). Participants’ roles (Table 1), level of education (Table 2) and time since completion of training (Table 3) varied. The answers to statements relating to diabetes knowledge, confidence and experiences of diabetes care are given in Figure 2. There were significant correlations between:

• Confidence in level of diabetes knowledge and confidence in delivering safe diabetes care  (t_b=0.640; P<0.01).
• Experience delivering diabetes care and confidence delivering safe diabetes care (t_b=0.570; P<0.01).
• Confidence in delivering safe diabetes care and knowledge of insulin therapy (t_b=0.534; P<0.01).
• Confidence in delivering safe diabetes care and confidence in liaising with the diabetes specialist nursing team (t_b=0.426; P<0.01). 
• Familiarity with the Trust’s management of hypoglycaemia guidelines (TMHGs) and knowledge of insulin therapy (t_b=0.256; P<0.05).
• Familiarity with the TMHGs and delivery of safe diabetes care (t_b=0.367; P<0.01). 

Forty-six participants (66.6%) agreed or strongly agreed that they were familiar with the TMHGs and 51 (75%) correctly considered hypoglycaemia to be a blood glucose level <4 mmol/L (see Table 4). There was no significant difference between familiarity with TMHGs and correct identification of hypoglycaemia (P=0.229). 

Participants were invited to select all appropriate treatments for hypoglycaemia from a list of five options: 67 (97.1%) selected glucose drink, 22 (31.9%) a chocolate bar, 21 (30.4%) tea and biscuits, and 41 (59.4%) orange juice. None selected “don’t know”. For further analysis participants were divided into two groups: those that had correctly identified hypoglycaemia treatments and those that had not. No statistically significant difference was found when the groups were compared to the two groups created based on level of familiarity with the TMHGs (P=0.798), see Table 5.

When invited to select all of the oral diabetes treatments that could cause hypoglycaemia from a list of metformin, gliclazide, pioglitazone, sitagliptin, metformin M/R and “don’t know”, 10 participants (14.5%) selected “don’t know”. Forty-eight (69.6%) correctly selected gliclazide. 

Asked what they would do if they had a patient successfully treated for hypoglycaemia who was due insulin with breakfast, 40 (60.6%) participants correctly indicated they would administer the insulin. Of the remainder, two (2.9%) would omit the insulin injection, one (1.5%) administer half the prescribed dose, 21 (31.8%) would request medical advice and one (1.5%) did not know what to do. When divided into two groups (correct and incorrect answers) and compared to perceived familiarity with TMHGs (those who agreed/strongly agreed vs neutral/disagreed/strongly disagreed), no significant difference was found (P=0.109). There was no significant difference when comparing correct/incorrect answers to confidence in knowledge around insulin therapy, see Table 6 (P=0.443).

Demographic data on age, gender, first language, country of training, role, time since completion of training, highest level of education and diabetes training in the past 3 years were collected. The statistically significant differences for demographic variables and diabetes knowledge are given in Table 7. Age, level of education, country of training and whether English is a first language impacted on the care given as a result of diabetes knowledge. 

Discussion
Diabetes knowledge among nurses
Significant gaps were identified among some participants in areas such as management of hypoglycaemia, oral diabetes treatments that can cause hypoglycaemia and management of insulin therapy. Not all participants were aware of the latest best practice. These findings correlate with other studies (Livingston and Dunning, 2010; Strider and Phillips, 2011; Ahmed et al, 2012; Modic et al, 2013). While all of the demographic data yielded statistically significant differences for at least one variable, time since completion of training (experience), level of education, diabetes training and role do not show a marked significant difference across diabetes knowledge (see Table 7). For example, when the variable for the scenario where a patient who has successfully been treated for hypoglycaemia and is due insulin was compared with the demographic variables, the only statistically significant difference observed was for country of training (P=0.008); all other variables were P>0.05.  Modic et al (2014) found that level of education and experience did not impact on diabetes knowledge and Gerard et al (2010) found no relationship between level of education and knowledge. 

Perceived versus actual knowledge
A theme that arose from the literature review was “perceived versus actual knowledge”, and this was explored in the questionnaire. Of the 45 participants who agreed or strongly agreed that they were familiar with the TMHGs (see Table 4), 11 (25%) incorrectly identified the hypoglycaemic range. Familiarity with the guidelines did not have a significant impact on the correct identification of hypoglycaemic level.

Discrepancies were also found regarding hypoglycaemia treatment and familiarity with the TMHGs (Table 5), the management of insulin therapy after hypoglycaemia and confidence in knowledge of insulin therapy (Table 6), and in the management of insulin therapy after an episode of hypoglycaemia and familiarity with the TMHGs. Continuation of basal insulin while on variable rate intravenous insulin infusion (VRIII) and confidence in knowledge of insulin therapy was another area of discrepancy.

When asked whether or not to continue basal insulin while a patient is on VRIII, 33 (47.8%) answered “yes”, 17 (24.6%) “no”, and 19 (27.9%) “don’t know”. No significant difference was found when comparing the variable to confidence around insulin therapy knowledge (P=0.342).

A statistically significant difference was found for the training sessions attended (P=0.018). This was one of the only areas of diabetes knowledge captured where attendance of training showed a difference. This and the high numbers of people who answered “don’t know” may reflect that continuation of basal insulin while on VRIII is a more recent area of guidance (Joint Diabetes Societies for Inpatient Care, 2014; 2016).

These findings highlight that while some nurses feel confident in their knowledge around certain areas of diabetes care, their answers to diabetes knowledge and care questions do not reflect this. There appears to be a marked discrepancy between perceived and actual knowledge. As there was a significant relationship between those who were confident in their level of diabetes knowledge and those who were confident in delivering safe diabetes care, these gaps in actual knowledge are potentially impacting on patient care.

It is worth considering which other factors impact on nurses’ diabetes knowledge and the transfer of knowledge into practice. Holmes and Dyer (2013) identify lack of confidence, peer influence, busy working environments and reluctance to challenge peers’ practice as factors that can contribute to this. Graue et al (2010) identify barriers to the transfer of knowledge into practice, such as fear of change and of negative judgements, competing demands, structural and organisational problems, and workplace environment. 

Continued education in diabetes
Nurses have a responsibility to ensure their knowledge and skills are up to date in order to practise safely (Nursing and Midwifery Council, 2015). Although the majority of participants (68.7%) had attended at least one diabetes training session in the past 3 years, a significant proportion had not. Considering the increasing prevalence of diabetes, complexity of diabetes management and the rapid changes in management practices, up-to-date diabetes knowledge is essential (Gerard et al, 2010; Ahmed et al, 2012; Modic et al, 2014). If RNs are unable to identify their own lack of diabetes knowledge and are unaware of the fast pace of changes in management, the relevance to them of continued education in diabetes may be diminished. 

Study limitations
Despite careful consideration and planning, potential response bias and social desirability cannot be dismissed (Bowling, 2014). A poor response rate is characteristic of questionnaires (Oppenheim, 1992) and although several efforts were taken to maximise responses, the sample of respondents falls short of the intended 86 participants. There is the potential for non-response bias (Bowling, 2014) where the characteristics of non-respondents differ significantly from those of respondents. While the sample cannot claim to be representative or enable generalisation, this project enabled the views and perspectives of 69 RNs to be captured and contributes towards the service evaluation. 

Changes made following the study
This project led to a reflection on current diabetes education practices, in particular considering how the training offered can address the knowledge gap. The awareness of a discrepancy between perceived and actual knowledge as well as the importance of culture and leadership led the Trust’s Director of Nursing and Head of Education to create a team dedicated to reviewing education practices. This involved the development of a diabetes competency workbook, practical workshops to facilitate transference of knowledge into practice, and the greater involvement of nursing leadership and the education and development team in diabetes training. 

Conclusion
There were significant gaps in diabetes knowledge among some RNs in the Trust. Not all participants were aware of the latest best practice. A discrepancy was found between perceived and actual knowledge. With the increasing prevalence of diabetes, its complexity and changes in management, it is essential RNs have current diabetes knowledge. 

Acknowledgement
This article is based on research undertaken at the Institute of Health and Society at the University of Worcester in 2015 for Christina Lange’s master’s degree dissertation. The dissertation was submitted and completed in July 2016.