Patients with end-stage kidney failure (ESKF) depend on safe chronic haemodialysis therapy to restore homeostasis. Safe haemodialysis treatment includes correct dosages of unfractionated heparin (UFH), currently the most widely used anticoagulant for this purpose. Failure to adhere to correct dosages of UFH exposes patients to the risk of blood clotting in the extracorporeal circuit or prolonged time for arterio-venous fistulae cannulated sites to stop bleeding (Davenport 2009) at the conclusion of the session. Renal unit practitioners need specialised knowledge and skill pertaining to the use of UFH in haemodialysis. Surprisingly, there is a dearth of available current literature on renal unit practitioners’ knowledge and practice regarding the safe administration of UFH, and even less available literature on practitioners’ attitudes concerning safe administration of UFH.

The primary purpose of the study was to describe renal unit practitioners’ knowledge, attitudes and practice (KAP) regarding the safe use and effects of UFH in selected adult, chronic haemodialysis centres in the Cape Town Metropole, South Africa. The secondary purpose was to determine whether there was an association between selected independent variables (category of renal unit practitioner, years of experience, duration of orientation period and in-service education in pharmacological actions of UFH) and practitioners’ KAP concerning the safe use and effects of UFH. The null hypotheses stated that the independent variables did not influence practitioners’ KAP regarding safe use and effects of UFH.

UFH has been used since the inception of chronic haemodialysis therapy in the 1930s. For patients on chronic haemodialysis, a nephrologist prescribes treatment for about 3–4 hours two or three times a week. During the haemodialysis, 200–250 mL/minute of the patient‘s blood volume is extracorporeal and can clot, but to reduce the risk of clotting (Roy & Kalra 2012:107) an anticoagulant such as UFH is administered. In local renal settings the South African Renal Society guidelines (2006) and other international guidelines are provided on the use and dosage of UFH.

Although 184 related published studies between 2000 and 2014 were located in three databases, no KAP studies related to renal practitioners’ use of UFH appear to have been published. One systematic review by the Cochrane Renal Group (Palmer et al. 2014) was located but was unrelated to the present study. Renal unit practitioners are reported to be confused, concerned and even outspoken about safety aspects of UFH when preparing, administering and monitoring its effects (Pittard 2001:75; Baglin et al. 2006:21; Brunet et al. 2008:794). Suranyi and Chow (2010:386) have appealed to nurses to regularly review and update their knowledge regarding UFH.

UFH prevents clotting in the extracorporeal haemodialysis circuit, but decades after its first use warnings continue of it being a high-alert medication (Institute for Safe Medication Practices [ISMP] 2006a) If safety precautions such as ensuring the right dose of UFH at the right time of administration and close monitoring are absent during haemodialysis therapy, serious adverse consequences such as anaphylaxis, heparin-induced thrombocytopenia and vascular access bleeding can result (Pittard 2001:75; Furushahi et al. 2002:1457; Bircher et al. 2006:1437). In the United Kingdom (UK), the National Patient Safety Agency (NPSA 2006) reported that anticoagulants are often the cause of preventable harm to patients, which can result in admission to hospital and even death. The Joint Commission on Accreditation of Healthcare Organisations (JCAHO) in the United States of America (USA) developed a project supporting the adoption of safety measures for anticoagulants as a National Patient Safety Goal, implemented in January 2009 (JCAHO 2009).

Approximately 35% of UFH is filtered by normal kidneys, but for patients on chronic haemodialysis, excess heparin cannot be excreted as it cannot be dialysed (Pittard 2001:75). Most protocols for use of UFH in the participating Cape Town Metropole dialysis centres prescribe a standard dose of 5000 IU of heparin three times a week for patients during haemodialysis sessions. This means that a patient's yearly dose of UFH amounts to approximately 780 000 IU, thus, predisposing the patient to long-term effects of UFH, such as osteoporosis (Pittard 2001:75; Furuhashi et al. 2002:1457). Although anticoagulants are widely used, errors and lack of tight control of the treatment of patients receiving anticoagulants continue (Institute for Healthcare Improvement [IHI] 2008) and heparin ‘is frequently unmonitored, untested and unsupervised in many dialysis facilities’ (Pittard 2001:75).

The NPSA reported 480 cases of patient harm from the use of anticoagulation therapy in the UK between 1990 and 2002 (NPSA 2006), of which 28 deaths were associated with the use of UFH but not linked to deaths of patients on chronic haemodialysis. Errors related to UFH accounted for 2.1% of total records submitted to the USA MedMarx national error database, of which 4.5% – 5.5% were harmful (Niccolai et al. 2004:146S). In a USA study (Grissinger et al. 2010:195), patient harm from UFH accounted for 1.4% to 4.9% of the reports submitted, occurring mostly (56%) during administration. Policy guidelines on monitoring baseline clotting studies before the administration of the initial and subsequent doses of UFH were available in only one of the seven Cape Town dialysis centres included in the study. In the other six centres, nephrologists prescribed individual patient dosages of UFH on medication charts. Five of the seven dialysis centres had a protocol on how to administer UFH. No data on adverse events (AEs) or deaths of adult patients on chronic haemodialysis related to UFH were available from the Cape Town Metropole. Moreover, there is no published data on morbidity and mortality associated with anticoagulation therapy within the South African renal population.

The results of this local study fill a gap in existing published data on renal practitioners’ KAP concerning the safe use of UFH, which may contribute to some extent to evidence-based renal nursing practice in South Africa to advance quality care and patient safety.

The study addresses gaps in the published literature concerning UFH prescription practices, dosage, administra­tion and monitoring of its effects (Ouseph & Ward 2000:181; Pittard 2001:75; Baglin et al. 2006:21; Brunet et al. 2008:794).

A descriptive cross-sectional design was employed for a KAP survey amongst renal unit practitioners (RNs, ENs and CTs) between March and April 2013.

Seven adult dialysis centres (Table 1), from two tertiary government hospitals and five dialysis centres from a private dialysis provider, were invited to participate.

A purposive sampling method was employed. To estimate a sample size in the absence of published data, the prevalence of a satisfactory level of knowledge of the safe use and effects of UFH amongst respondents was used and set at 50% (n = 52) of the population (N = 104), guided by personal clinical experience. Using StatCalc (EpiInfo Version 7), a sample size of 82 was determined. Although attitudes and practice scores were also investigated, these variables were not used to estimate a sample size.

All recruited dialysis centres participated in the study (Table 1): two tertiary government hospital dialysis centres (28.6%) and five dialysis centres (71.4%) from a private dialysis provider. Of the total population (N = 104) of renal unit practitioners who met inclusion criteria, 77 (74.0%) participated in the study compared to the estimated sample size (n = 82, 93.9%), therefore, no questionnaires were excluded from the final data analysis, even those with missing data. The highest proportion of missing data (8.2%) was recorded for the knowledge section. There was a 70.0% response rate (35/50) of practitioners from tertiary hospitals, and from the private sector, it was 77.8% (42/54).

The majority of respondents (94.8%, 73/77) were South African citizens (Table 2). There was a statistically significant association (P = 0.002, SD±10.8) between the variable professional category and age. A Scheffé post-hoc test showed that the difference in age between RNs and CTs reached statistical significance (P < 0.001, 95% CI 7.56–19.41) and between ENs and CTs (P = 0.020, 95% CI 1.70–24.24), but not between RNs and ENs. The association between gender and professional category reached statistical significance (P < 0.001). There was a statistically significant association (P < 0.001) between the variable professional category and qualification. Most respondents (53.2%, 41/77) had 0–5 years of experience (mean 9.4, SD ± 9.3), which did not reach statistical significance for the professional category. There was a statistically significant association (P = 0.001) between the variable professional category and having or not having in-service education on the pharmacological actions of UFH.

Respondents’ overall test scores achieved for KAP are presented in Table 3.

The mean score for knowledge was 10.8/18 (60.0%); for attitude, 25.9/41 (63.2%); and for practice, 35.8/50 (71.6%).

Results for specific questions on respondents’ knowledge of the use and effects of UFH are presented in Appendix 1 Table 1 - A1.

The association between knowledge and selected independent variables is presented in Table 4.

The data in Table 4 shows that overall, 75.3% (58/77) of respondents had an acceptable level of knowledge regarding UFH, whilst 24.6% (19/77) had poor knowledge. Most of the RNs (82.3%, 42/51) had an acceptable level of knowledge. The majority (80.0%, 4/5) of ENs had poor knowledge of UFH, despite the fact that as employees within haemodialysis units, they administer UFH under supervision of RNs. There was a statistically significant association (P = 0.011) between the variable professional category and self-reported knowledge of UFH. Bivariate logistic regression analysis showed that the odds of the ENs having poor knowledge regarding UFH compared to the RNs were 18.7 times higher at a 95% CI (1.9–187.4), and that this difference reached statistical significance (P = 0.013). The odds of the CTs having poorer knowledge than the RNs were 1.9 times higher, at a 95% CI (0.6–6.1), but this difference did not reach statistical significance (P = 0.304).

The majority of respondents 41.5% (32/77) who had an acceptable level of knowledge of UFH had 0–5 years of experience. Ten of the 77 respondents (12.9%) who had 11–35 years of experience displayed poor knowledge of UFH. There was no statistically significant association between the level of knowledge and the variables:

years of experience (P = 0.604)

For in-service education on UFH, 1.3% (1/77) of respondents were not included in the calculation.

Data showed that the respondents’ lack of knowledge concerning the use and administration of UFH was reflected in their attitude to its use and administration. Data in Table 5 shows the association between attitudes and selected independent variables.

Data in Table 5 show that most RNs (74.5%; 38/51), CTs (66.6%; 14/21) and ENs (60.0%; 3/5) had a positive attitude, but the association between professional category and attitude did not reach statistical significance. The association between years of experience and attitude and between in-service education on UFH and attitude was not statistically significant.

Results for specific questions on respondents' practice regarding the use and effects of UFH are not presented in this article, but are available on request from the authors. Data in Table 6 show the association between practice and selected independent variables.

Across the different professional categories of renal unit practitioners, the results reveal that most had an acceptable level of self-reported competency regarding the use of UFH. The association between the independent variable professional category and the dependent variable practice was not statistically significant (P = 0.517).

Of the respondents with 0–5 years’ experience, most (63.4%, 26/41) reported an acceptable level of practice competency. The association between years of experience and self-reported practice concerning UFH was statistically significant (P = 0.016), therefore, bivariate logistic regression analysis was performed for strength of association between the two variables and for statistical significance, before making further judgements about the null hypothesis. Respondents who had 0–5 years of experience were 4.6 times more likely to report having poor practice at a 95% CI (1.4–15.6), than respondents who had more than six years of experience, and this difference reached statistical significance (P = 0.014).

There was no statistically significant association between duration of orientation and self-reported good practice (15/20, 75.0%) in relation to UFH. The results of the Fisher's exact chi-squared test revealed that the association between in-service education and practice reached statistical significance (P = 0.028), and bivariate logistic regression analysis was performed. The odds of respondents who had no in-service education reporting poor practice were 4.3 times higher, at 95% CI (1.1–16.5), than respondents who had education in pharmacological actions of UFH, and this difference reached statistical significance (P = 0.032).

The outcomes of the logistic regression tests confirm rejection of the null hypotheses that the variables professional category, years of experience and in-service education on the pharmacological actions of UFH do not influence renal unit practitioners’ KAP. However, the null hypothesis, that the variable duration of orientation did not affect the renal unit practitioners’ KAP, was accepted.

Results of a Malaysian KAP study on general medication amongst 40 medical nurses showed adequate knowledge (mean scores: knowledge 13.8, attitude 16.4 and practice 10.7), but a lack of in-depth knowledge of aspects of pharmacology (Raja, Daud & Syed 2009:17). The mean score for nursing students’ knowledge of pharmacology in 29 nursing schools was 55% and for medication calculation skills 66%, but they performed poorly in knowledge retention and practical implementation in clinical areas (Dilles et al. 2011:499). In our study the majority of RNs and CTs had an acceptable level of knowledge of UFH but, overall, respondents reported incorrectly that UFH breaks down blood clots, whereas it prevents blood clots forming (Fischer 2007:178; Lankshear, Harden & Simms 2010:49), as UFH is not a fibrinolytic agent. Most respondents did not know the correct antidote to or dosage for managing UFH overdose (1 mg of protamine sulphate to neutralise 100 IU of UFH [European Best Practice Guidelines Expert Group on Haemodialysis 2002:65; Baglin et al. 2006:28; Lankshear et al. 2010:51; Lemon & Crannage 2011:212]). Most respondents did not know the reason for not administering UFH orally (it is poorly absorbed). Most respondents did not know that the type of heparin used in their unit was a porcine derivative which carries a greater risk of anaphylaxis (Shen & Winkelmayer 2012:475)

Some respondents’ inability to calculate correct dosages of UFH was of deep concern, as this can have life-threatening consequences for patients. Nurses’ medicating deficiencies were highlighted in a study conducted on medication knowledge, dose calculation (Barkhouse-Mackeen & Murphy 2013:91) and medication errors which resulted from poor calculation skills (Polifroni, McNulty & Allchin 2003:458; Wright 2007:279). A study on numerical skills and drug calculations in relation to patient safety, conducted in the UK in 2006 (McMullan, Jones & Lea 2010:891) showed, that 55% (126/229) of nursing students and 45% (20/44) of RNs failed the numeracy test, and 92% (211/229) of the students and 89% (39/44) of RNs failed the calculation test.

Respondents’ lack of knowledge in certain areas was reflected in their attitude to the use and administration of UFH, but extrapolation to mean that poor knowledge translates to having poor attitudes must be interpreted with caution. Only a few respondents were greatly concerned that patients may experience adverse effects of UFH, implying a casual attitude regarding such a high-alert medication. More than half of the respondents felt uncomfortable administering UFH that had been prepared by others, and this attitude to practice is supported in the published literature (Chevalier et al. 2011:344).

The majority of respondents displayed a positive attitude in relation to the administration of UFH. Respondents with 0–5 years of experience (65.9%; 27/41) presented the most positive attitude.

The majority of all categories of renal practitioners reported having acceptable practice. The majority would always give the UFH dose prescribed by the nephrologist, and about half of them would contact the prescribing doctor if concerned about a patient's prescribed dose, whilst a few would adjust the dose and document the adjusted dose. Rather than first consulting the medicines formulary for UFH information, respondents would consult a doctor. Most respondents would always monitor patients for anaphylactic shock and few would never verify allergies with patients before administering UFH. The majority of respondents would always check the UFH dose with another person before administration, rather than only doing so sometimes. Most respondents would always document adverse effects of UFH and report related adverse effects in patients to unit managers. A study by Hanafi et al. (2012:21) in Tehran showed that 91% of nurses had never reported an adverse drug reaction (ADR) directly to the Iran Adverse Drug Reaction National Centre and 35% were unaware of the centre's existence, but were most likely to report to doctors (87.1%) and pharmacists (1.8%). The reasons for not reporting ADR were:

uncertainty of the adverse reactions (49.5%)

However, Niccolai et al. (2004:148S) suggest that non-reporting directly to ADR databases could be attributed to nurses fearing reprisal and blame. A multinational systematic review of empirical evidence on the prevalence and nature of medication administration errors (Keers et al. 2013:1) showed that, despite improvement in medication safety, higher medication administration error (53.3%) occurred via the intravenous route compared to other routes.

Study findings suggest that currently, in the seven research sites, there is not a set schedule for coagulation studies for patients on haemodialysis. Pittard (2001:75) and Winkler et al. (2007:499) support the need for strict laboratory monitoring of coagulation studies of patients, to establish a therapeutic dose to prevent under or over-anticoagulation. Respondents did not consider prolonged bleeding or clotting of patients’ blood circuits (Pittard 2001:74) to be the adverse effects of UFH, which may account for them reportedly not having recently witnessed an adverse event. Respondents had not kept up to date with literature specifically on UFH, a concern echoed by Ndosi and Newel (2008:570) and Raja et al. (2009:17) that, in general, nurses lack in-depth knowledge of medication.

This study was evaluated using The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) (Vandenbroucke et al. 2007) checklist, to ensure that all the important aspects of a descriptive cross-sectional study were attended to.

The limited statutory scope of practice for ENs (South African Nursing Council, Republic of South Africa 1984) was reflected in their poor knowledge of UFH, and this has implications for their placement in haemodialysis units. Therefore, ENs should not practise independently in a haemodialysis unit, but at all times under the supervision of a registered nurse and not a clinical technologist. Pre and post-registration nursing education programmes should emphasise medication calculation, pharmacokinetics, pharmacotherapeutics and pharmacodynamic aspects of UFH. Nurses should be taught how to navigate medicine formularies and how to do literature searches to improve their medication knowledge to ensure life-long learning.

Guidelines for safe administration of UFH and monitoring practices should be developed through consensus by an expert panel. Management teams should implement problem-based, in-service education programmes that include pharmacology, pharmacokinetics, pharmacodynamics, medication calculation and pharmacotherapeutics. Improving staff KAP regarding UFH should limit medical litigation and financial claims against individual health care personnel and institutions. Management teams should encourage renal unit practitioners to enrol for patient safety programmes such as the webinar by ISMP (2013:1), ‘Improving Medication Safety through Staff Education and Competency Assessment: An Important Challenge for Healthcare Organisations’.

Management teams should establish pharmaco-vigilance units to encourage reporting of incidents and unsafe practice without naming, blaming and shaming, to augment the renal database of patients who experienced AEs associated with UFH. Staff could be provided with a bedside calibrated point-of-care testing apparatus to measure patients’ activated partial thromboplastin time instead of having to wait for laboratory results.

Before new staff administer UFH, their competency level should be evaluated before and after the orientation programme and, if they do not pass the predetermined set score, remedial intervention should be instituted. Staff competency tests should be conducted annually and if results are poor, re-training and re-testing should be offered to ensure ongoing employment.

Shen and Winkelmayer (2012:483) suggest that large cluster randomised trials should be conducted in haemodialysis units to confirm the safe therapeutic range for different anticoagulants or pertaining to their risks and benefits. A renal nursing practice model should be developed for the local South African context for both the public and private chronic haemodialysis centres.