Improving peripheral venous cannula insertion in children: a mixed methods study to develop the DIVA key

Insertion of a peripheral intravenous catheter (PIVC) is almost synonymous with hospitalisation [1, 2]. As a vascular access device, it is minimally invasive and facilitates immediate medical treatment. However, most children and their families describe insertion of a PIVC to be one of the most painful and stressful procedures during their hospitalisation [3]. Up to 69% of first attempt insertions fail [4‐6], leading to delays to medical treatment and extended inpatient days [7, 8]. For health services, PIVC insertion failure contributes to significant usage and wastage of healthcare resources, costing the Australian health care system nearly $450 million [AUD] annually [9].

More than 50% of children are conservatively estimated to have difficult intravenous access (DIVA) [5, 10, 11]. DIVA is characterised by nonvisible and non-palpable veins; which may be due to physiology, pathology, or previous PIVC damage; making PIVC insertion ‘difficult’ for most clinicians [6]. The ‘average’ PIVC insertion requires two attempts and 20–30 min [12]. For children with DIVA, successful PIVC insertion may require upwards of nine attempts (needle sticks) [5]. Historically, DIVA status is retrospectively assigned after the patient has endured multiple failed PIVC insertion attempts. As a result, there has been a recent surge in the development of DIVA decision-making resources for paediatric patients [13]. Processes to identify children with DIVA have been attempted, primarily within emergency departments [14] with the development of three [11], four [10, 11, 15] and five variable DIVA Scores [16]; a Peripheral Venous Grading System [17] and a Peripheral Vein Assessment Instrument [18]. While many have been based on sizeable cohorts, their clinical generalisability in general hospital wards can be limited [10, 11, 15]. Furthermore, they are i) limited in their capacity to direct clinicians on how to manage these ‘DIVA’ patients once identified; and ii) typically do not feature consumer engagement, or patient centred principles such as procedural pain, or skill and confidence of the inserter. Together with input from key clinical stakeholders and pediatric consumers, we sought to develop a DIVA identification and escalation instrument (the DIVA Key) to support clinical practice. The reliability, reproducibility, and acceptability of the DIVA Key, as a strategy to identify children with DIVA, was then evaluated to assess clinical utility and application feasibility in paediatric settings.

A sequential, mixed methods study was undertaken at the Queensland Children’s Hospital (QCH) Australia, a quaternary paediatric hospital, between February and December 2020. Study design was underpinned by the Co-production and Prototyping framework for Public Health Interventions [19]. Ethical approval was obtained from Children’s Health Queensland Human Research Ethics Committee (LNR/19/QCHQ/55326) and Griffith University (2019/797). The study is reported in line with The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement [20] and informed by the COnsensus-based Standards for the selection of health status Measurement INstruments (COSMIN) checklist [21].

This project co-produced and validated the DIVA Key instrument to support the assessment and subsequent escalation of PIVC insertion care in children with DIVA. Consistent with the evidence-base around DIVA risk factors and existing DIVA instruments [10, 11, 13‐18], the DIVA Key included objective descriptions of vessel quality (e.g., appearance) and practice variables (e.g., acuity, previous access history). These indicators were uniquely complemented by rating the child’s reported or perceived anxiety, along with the clinician’s self-assessment of ability, and recommendations for management of procedural pain and anxiety. In this study the DIVA Key demonstrated interrater reliability and agreement and was acceptable to both clinicians and families. Our findings support the instrument’s application and utility across a quaternary paediatric hospital setting.

Our prior literature review and survey of practice [13, 40] demonstrated a variability surrounding the sensitivity and useability of paediatric DIVA instruments. Whilst most tools demonstrated moderate predictive ability (DIVA Score area under the curve [AUC] 0.67 75% [15], 3-variable DIVA score AUC 0.72 [11]), clinometric testing outside of the emergency department was scarce. Further existing DIVA tools lacked decision-making cues to direct escalation, failed to consider inserter skill and confidence, and did not take into consider patient experience or preference (i.e., patient reported pain and anxiety). To overcome these obstacles, the DIVA Key was co-produced with stakeholders and consumers, and underpinned by the diverse and unique insights they provided [41]. The resulting resource, the DIVA Key, reflects the experiences of clinicians and consumers, and demonstrates promising reproducibility, utility, and acceptability. Due to the insights from stakeholders and consumers, the DIVA Key had strong face and content validity, and stakeholders qualitatively reported that they felt their feedback was incorporated into the final instrument.

PIVC insertion in patients with DIVA is challenging, even for experienced clinicians [42]. Irrespective of inserter skill, identification of DIVA risk should occur prior to subjecting a child to multiple, painful failed insertion attempts [43]. Early identification of children with DIVA relies on instruments that, even with minimal training, provide an objective and reproducible description of a child’s risk of DIVA. To achieve this, the DIVA Key employed alliteration ( [44] drawing emphasis to certain ideas) to prompt recall by users of DIVA risk factors (acuity, appearance, alerts, admissions, age, and anxiety) and a traffic-light system [45] for risk level, complemented by the clinician self-assessment (ability). As a result, there was considerable agreement among assessors of varied skill level when assessing the child’s predicted risk of DIVA, and a high level of inserter-reported satisfaction with the instrument. Additionally, the DIVA Key demonstrated high construct validity. Children who were assessed as ‘high risk’ had the lowest rate of first-attempt insertion success, slightly higher median insertion attempts, and greater perceived insertion difficulty.

Early identification of DIVA in children alone is not sufficient to improve patient outcomes. Clinicians in this study voiced concerns that they lacked the skills, training, or resources to manage patients with DIVA. This finding is consistent with earlier research highlighting the lack of support or resources available to clinicians once DIVA is identified [13]. Therefore, the DIVA Key includes a clear and concise escalation pathway (ascend), that matches patient level of risk and inserter competency, to guide the number of attempts before escalation (i.e., no more than 2 attempts prior to escalating to a more experienced clinician with or without vessel visualisation technologies). Similar to Hallam’s vessel preservation tool [18, 46].

This study has several strengths. Instrument development was grounded in a literature review, survey of practice and extensive stakeholder consultation [47, 48]. Additionally, the instrument development was underpinned by the co-production framework [19] and develop iteratively which allowed for constant adaption and improvement based on key stakeholder requirements. Despite these strengths our study is not without limitations. The development of the DIVA instrument relied heavily on clinical expertise. Although this level of evidence can be inconsistent, it was necessary to ensure the utility and acceptability of the instrument. Previous research demonstrated the complementary role of clinical ‘gestalt’ in accurate DIVA identification [29]. Finally, in this study, inserters were typically resident medical officers or registered nurses as is standard in our local setting. Insertion practices vary between institutions globally however, and interpretations of the concordance and utility of the instrument should be interpreted accordingly. The design of the current study precluded evaluation of the criterion validity of the DIVA Key, and therefore we were unable to determine the sensitivity and specificity, and corresponding positive and negative predictive values, of the DIVA Key in detecting a child’s DIVA status prior to escalation. Given the indicators of DIVA used in the DIVA Key are consistent with other highly sensitive paediatric instruments (e.g., Cornell Assessment of Pediatric Delirium [49, 50]), and the results of logistic regression found that first attempt insertion success was lowest in children assessed as high risk of DIVA, however, suggests that the DIVA Key is a promising instrument for the accurate identification of DIVA in this population.

We co-produced a user-friendly, consumer focussed instrument to support the identification of DIVA in children, with an inbuilt inserter escalation pathway. In this cohort the DIVA Key appears to be a reliable instrument to support the identification and management of DIVA in children. Further testing among varying cohorts is warranted to further test generalisability of this measurement instrument and determine whether the implementation of an escalation pathway improves patient outcomes.