An Improved Definition and SAFE Rule for Predicting Difficult Intravascular Access (DIVA) in Hospitalized Adults

Abstract

Patients with difficult intravascular access (DIVA) are common, yet the condition is often ignored or poorly managed, leading to patient dissatisfaction and misuse of health care resources. This study sought to assess all published risk factors associated with DIVA in order to promote prospective identification and improved management of patients with DIVA. A systematic literature review on risk factors associated with DIVA was conducted. Risk factors published in ≥4 eligible studies underwent a multivariate meta-analysis of multiple factors (MVMA-MF) using the Bayesian framework. Of 2535 unique publications identified, 20 studies were eligible for review. In total, 82 unique DIVA risk factors were identified, with the 10 factors found in ≥4 studies undergoing MVMA-MF. Significant predictors of DIVA included vein visibility, vein palpability, history of DIVA, obesity (body mass index [BMI] >30), and history of intravenous (IV) drug abuse, which were combined to create the mnemonic guideline, SAFE: See, Ask (about a history of DIVA or IV drug abuse), Feel, and Evaluate BMI. By recognizing patients with DIVA before the first insertion attempt and treating them from the outset with advanced vein visualization techniques, patients with DIVA could be subject to less frequent painful venipunctures, fewer delays in treatment, and a reduction in other DIVA-associated burdens.

INTRODUCTION

Approximately 70% to 80% of hospitalized patients will require a peripheral intravenous catheter (PIVC) during their hospital stay,1 of whom up to 59.3% are classified as having difficult intravascular access (DIVA).2 Unfortunately, there has been no clear consensus on what defines DIVA. A recent systematic literature review (SLR) proposed a new, evidence-based definition of DIVA: “when a clinician has two or more failed attempts at peripheral intravascular (PIV) access using traditional techniques, physical examination findings are suggestive of DIVA (eg, no visible or palpable veins) or the patient has a stated or documented history of DIVA.”3 Including failed attempts in the definition is problematic, as patients continue to be subjected to multiple unnecessary venipunctures. Thus, there is a clear need to better characterize patients with DIVA preemptively and to adjust the definition accordingly.

The excessive burden of multiple PIVC attempts for patients with DIVA should not be overlooked, as the impact extends not only to the patients, but also to the clinical staff and the health care system as a whole. Placement of a PIVC for patients with DIVA can be very time consuming, with substantial delays potentially leading to significant treatment delays. Frequent PIVC attempts can also exacerbate venous depletion and may be painful and anxiety inducing.4 Further, multiple attempts can lead to a significant increase in costs and be potentially hazardous for the health care worker.4 Fortunately, there are more advanced methods of venous access that can assist the inserter and increase first venipuncture success.5 These techniques may help reduce the burdens associated with multiple PIVC insertion attempts.6 However, they require additional equipment, supplies, and expertise and, therefore, are not practical to implement in patients without a clear indication.7 Without a practical identification of a patient with DIVA and clear indication for advanced access, patients endure multiple PIVC insertion attempts, additional supplies utilization, and need for clinical expertise, leading to increased burden. Thus, it would stand to reason that the optimal DIVA management strategy would involve screening all patients prior to any attempted PIVC insertion by standard techniques.

Recent SLRs have identified an abundance of screening rules, tools, or algorithms for predicting DIVA, but, to the authors' knowledge, none of these prediction devices have been widely adopted in practice.8–10 The existing tools vary in the risk factors they include, but there are some factors that are nearly unanimously considered to be predictors of DIVA, yet none of the existing instruments consider all of the available data via evidence synthesis. A DIVA prediction tool developed using evidence synthesis techniques permits the prospective identification of patients with DIVA using the most complete set of information possible.

The present SLR considers evidence from all available DIVA risk factor studies in a meta-analysis, highlighting the most consistently identified risk factors for predicting DIVA. Additionally, a supplemental review compared the existing tools for predicting DIVA with the risk factors identified in the present SLR and meta-analysis. The goal of the analysis was to synthesize the key DIVA risk factors to promote an improvement in the current standard of care and to reconsider the multiple venipuncture paradigm for patients with DIVA. These changes have the potential to improve patient experience and decrease health care resource use by reducing the number of repeated attempts at venous access. Based on the information presented above, there is a clear need for improving the first-time insertion success rates for patients with DIVA.

METHODS

Literature Review

An information specialist searched the following databases using PRESS reviewed search terms on March 9, 2022 (see Appendix 1, http://links.lww.com/JIN/A110): Ovid MEDLINE(R), Ovid MEDLINE Epub Ahead of Print, In-Process & Other Non-Indexed Citations and Daily, Ovid EMBASE, EBM Reviews - Cochrane Central Register of Controlled Trials January 2022, EBM Reviews - Cochrane Database of Systematic Reviews 2005 to March 9, 2022, EBM Reviews - Database of Abstracts of Reviews of Effects 1st Quarter 2016, EBM Reviews - Health Technology Assessment 4th Quarter 2016, and EBM Reviews - NHS Economic Evaluation Database 1st Quarter 2016. The search returned 2535 unique results.

Studies were selected for the review if they included hospitalized or emergency room patients aged 18 years or older with an indication for receiving a peripheral IV, with or without ultrasound guidance. Only English-language studies were considered, and studies were required to report risk factors associated with DIVA. Publications that did not report risk factors, used only central venous catheters, reported on patients in the prehospital setting, or reported retrospective or survey data without the inclusion of a DIVA prediction tool were excluded. Study selection was performed in duplicate by 2 reviewers, and details are illustrated in the PRISMA flow diagram (Figure 1). As most of the included reports were noninterventional, unblinded, prospective, and observational studies, standard risk of bias tools could not be applied; thus, risk of bias was assessed descriptively.

Figure 1.

PRISMA flow diagram.

Data were extracted by 2 reviewers to a predefined Microsoft Excel (Microsoft, Redmond, WA) spreadsheet that captured study details, IV characteristics, patient background, provider characteristics, and all reported risk factors (adjusted and unadjusted univariate or multivariate relative risks [RRs] and/or odds ratios [ORs] and 95% confidence intervals [CIs]). When ORs were not directly available, they were calculated based on the relevant contingency tables. When trials reported multiple outcome definitions, failure of the first insertion attempt was preferentially selected. Quality checks were performed by a third reviewer.

After risk factors were extracted from the included studies, synonymous terms, as identified by a clinical expert, were combined where possible. A clinician-identified set of a priori factors expected to be important in predicting DIVA was also considered. These a priori factors were lack of vein palpability, lack of vein visibility, obesity, history of DIVA, history of dialysis, and insertion site availability.

Additionally, a supplemental comparison with existing tools was conducted. ORs were considered against those reported in other similar prediction tools to check for alignment and to support the development of a guideline for clinical use.

Statistical Analysis

To the authors' knowledge, there is no evidence of a common instrument for prospectively predicting DIVA that has been widely adopted for clinical use.8 Most work has focused on the creation of novel methods for predicting DIVA and has resulted in multiple similar, yet not directly comparable, tools. To synthesize the available information, a meta-analysis of all the collected risk factors for DIVA was conducted. When estimating the association of multiple risk factors with an outcome, parameters were evaluated through a single multivariate meta-analysis of multiple factors (MVMA-MF), as well as individual univariate meta-analyses (UVMA).

Due to the paucity of data, each DIVA risk factor may have only 2 to 3 studies contributing direct evidence. When conducting individual UVMAs, each pooled risk factor may be using information from dissimilar study populations, which impacts the comparability and generalizability of the findings. To address this challenge to comparability and generalizability, a MVMA-MF was used to synthesize the data, attempting to reduce bias by using information from the correlations between multiple factors within a study. Due to the correlations within each study being unknown and a lack of available individual patient data (IPD), a hybrid MVMA-MF was specifically conducted.11

Only risk factors that were reported in at least 4 studies were included in the analysis, as including risk factors with even more sparse data prevented the model from converging. Each risk factor was also evaluated individually through UVMA as a sensitivity analysis. Many studies reported multivariate analyses; however, the adjustment sets varied greatly, so meta-analyses on adjusted ORs could not be conducted. The MVMA-MF analyses were conducted under the Bayesian framework using the altmeta R package.12 Random-effects models with vague priors were used. Models were run with 3 independent chains with at least 10 000 burn-in samples and at least 100 000 sample iterations. Model convergence was assessed using R-hat and visual inspection of trace plots. The UVMA analyses were conducted using random-effects models and the inverse-variance method in the metafor R package.13 Results were reported as ORs with credible intervals (CrI).

RESULTS

Literature Review

Of the 2530 studies screened, 96 were assessed at the full-text level and 20 studies were ultimately selected for inclusion. Of the 20 studies included in analysis, 9 were conducted in Europe, 6 in the United States, 3 in Australia, and 1 each in Turkey and Japan. Just over half of the studies (11 of 20) included patients who presented to the emergency department. Other studies took place in surgical wards (n = 3), hospital wards (n = 2), and a radiology department (n = 1). Two studies had a mix of patients from emergency departments and surgical or hospital wards, and 1 study was conducted across several hemophilia treatment centers. Fifteen of the 20 studies included exclusively adult patients, one of which was limited to patients over the age of 65 years. One study was open to patients of any age, and the remaining 4 did not specify any age restrictions. The studies had moderately heterogeneous definitions of DIVA, with failure after ≥1 attempt (13 studies) being the most common definition. Other definitions considered a history of DIVA (n = 1), the need for a rescue technique (n = 2), abandonment of the procedure (n = 1), no visible or palpable veins after placing a tourniquet (n = 1), procedure duration of more than 1 minute (n = 1), use of central venous access (n = 1), and failure after 2 or more (n = 1) or 3 or more (n = 2) insertion attempts (Table 1).14–31 The literature review identified DIVA risk factors for a total of 15 628 patients, but the heterogeneity of collected risk factors was a constraint on the sample size available for each of the factors in the current analysis.

TABLE 1. Study Location, Sample Size, Setting, Objective, Design, and Definition of DIVA for Included Publications.

First Author (year)	Country	Sample Size	Setting	Objective	Study Design	Definition of DIVA
Abe-Doi14 (2021)	Japan	32	Gastroenterological medicine ward	To investigate the effectiveness of ultrasound in improving the success rate of catheterization.	Prospective observational cohort	Failed first insertion attempt
Armenteros-Yeguas2 (2017)	Spain	135	Hospital	To estimate the prevalence of difficult venous access in complex patients with multimorbidity and to identify associated risk factors.	Cross-sectional	One or both of: (1) Personal history of DIVA, based on report or a record of more than 2 attempts, and (2) No visible or palpable veins after placing a tourniquet.
Carr15 (2016)	Australia	460	Emergency department	To identify the reasons for PIVC insertion in the emergency department, the first-time insertion success rate, and the patient and clinician factors influencing this phenomenon.	Prospective cohort	Failed first PIVC insertion attempt
Carr16 (2019)	Australia	879	2 tertiary emergency departments	To identify the incidence of, and factors associated with, first time PIVC insertion success in the emergency department.	Prospective cohort	Failed first-time insertion attempt
Civetta17 (2019)	Italy	1006	Hospital	To develop and validate a scale that is specifically designed to be applicable to all adult patients during the presurgical clinical examination, in order to identify patients at risk of peripheral DIVA.	Prospective cohort	Failed >3 PIVC insertion attempts
Fields18 (2014)	USA	743	Emergency department	To determine risk factors for DIVA in adult patients presenting to the emergency department.	Prospective observational	Failed ≥3 PIVC insertion attempts, or use of rescue technique required
Guillon19 (2015)	France	450	4 hemophilia treatment centers affiliated with university hospitals	To survey the frequency, causes, and clinical manifestations of DIVA and evaluate the clinical utility of a near-infrared vein visualizer.	Prospective	Not reported
Jacobson20 (2005)	USA	339	Large urban hospital	To determine which factors contribute to difficult PIVC insertion in hospitalized patients and are associated with IV insertion failure and success, and which special techniques nurses use to facilitate IV catheter insertion.	Descriptive	Failed PIVC insertion
Piredda21 (2017)	Italy	763	Radiology service of an acute university hospital	To identify risk factors for difficult intravenous cannulation in relation to characteristics of patients, health care providers, and devices in adult patients accessing a radiology service.	Prospective observational	Failed first attempt or more than 1 min needed
Rippey22 (2016)	Australia	734	Emergency department	To explore whether clinicians of differing experience levels can predict their own likelihood (clinician “gestalt”) of first-time PIVC insertion success on any given patient.	Prospective self-report	Failed first insertion attempt
Rodriguez-Calero23 (2020)	Spain	2662	Medical and surgical hospitalization wards, surgical areas, A&E and intensive care units, at 8 hospitals	To consider the risk factors, determine the influence of the hospital setting, to examine the association between DIVA and the different techniques of catheter insertion, and to describe the role of the clinician's experience in this context.	Case-control	Failed ≥2 attempts, use of rescue technique or central access, or abandonment of the procedure
Sebbane24 (2013)	France	563	Urban university hospital emergency department	To investigate DIVA across BMI categories in the emergency department, focusing on patient-related predicting factors.	Prospective observational	Failed ≥1 attempt
Shokoohi25 (2014)	USA	244	Emergency department	To determine whether patient and clinical factors successfully predict DIVA in adult emergency department patients.	Randomized controlled trial	Failed ≥1 attempt
Tran26 (2019)	USA	119	Emergency department	To compare standard elastic tourniquets with balloon pressure cuffs to promote venodilation and to assess how they influence first attempt PIVC insertion success.	Prospective, single-blinded, randomized controlled trial	Failed ≥1 attempt
Van Loon27 (2021)	Netherlands	660	Department of anesthesiology	To assess if inserting a notched peripheral intravenous catheter will increase first attempt cannulation success when compared to inserting a catheter without a notched needle.	Block-randomized trial	Failed ≥1 attempt
Van Loon28 (2021)	Netherlands	610	Hospital	To assess the association between catheter to vein ratio and the success rate of first PIVC insertion attempts.	Post-hoc analyses based on a previous observational study	Failed ≥1 attempt
Van Loon29 (2016)	Netherlands	1063	Tertiary hospital specialized in cardiothoracic, bariatric, and oncological surgery	To identify risk factors for failure to perform PIV cannulation on the first attempt in adult patients, to make it possible to calculate the risk of first attempt failure, and to prospectively classify patients with DIVA.	Prospective, observational, cross-sectional cohort study	Failed PIVC insertion on the first attempt
Van Loon1 (2019)	Netherlands	3587	5 hospitals	To improve the Adult Difficult Intravenous Access Scale (A-DIVA) by external validation and to predict the likelihood of DIVA in adults.	Multicenter and multidisciplinary cross-sectional	Failed ≥1 attempt
Witting30 (2012)	USA	107	Emergency department of an urban tertiary care university hospital	To estimate the incidence of DIVA and its associated delays in an urban emergency department.	Prospective cohort	Failed ≥1 attempt
Yalcini31 (2019)	Turkey	472	Emergency department	To determine the factors affecting the first-attempt success of PIVC placement in older patients in the emergency department.	Prospective, observational	Failed ≥1 attempt

Abbreviations: A&E, accident and emergency; BMI body mass index; DIVA, difficult intravascular access; IV, intravenous; PIV, peripheral intravascular; PIVC, peripheral intravenous catheter.

Patient Characteristics

Generally, reporting of patient baseline characteristics varied across studies. No characteristics were extensively reported, and some important characteristics, like race and dialysis status, were severely underreported, appearing in only 5 and 3 studies, respectively.

The distribution of mean values for key baseline characteristics is presented in Figure 2. Generally, patient characteristics across studies cluster around the median value, with the exception of some outliers. While patients in the majority of studies were predominately White (>75%), 1 study reported a population in which race was more balanced (40% Black, 48% White, 12% other – further detail not provided).18

Figure 2.

Kernel smoothed density plot of patient baseline characteristics reported in individual studies. Vertical black lines indicate median values across included studies. Characteristics are reported as percentage of patients with the trait, except for the characteristic of age, which is reported in years. Abbreviations: DIVA, difficult intravascular access.

Risk of Bias

All included studies were assessed for risk of bias. Most of the included studies were noninterventional and, thus, cannot be used to inform causation. The studies generally included any patients receiving a PIVC in the hospital or emergency department setting during the study period and can be expected to be an accurate representation of the hospital population, although the findings may not be as generalizable to individuals who are not hospitalized as frequently. Three of the included studies may be biased toward healthier patients, as they excluded critically ill patients.18,21,26 One study excluded patients with historical or current IV drug abuse, which may be a source of bias since other included studies found IV drug abuse to be an important risk factor for DIVA.24 Only one of the studies blinded participants, and none blinded providers, so there is a risk of implicit bias, especially in studies based on provider reports.26 As there was a lack of standardization in terms used to describe DIVA risk factors across studies, interpretation of the data may be impacted and may pose a challenge to developing a true understanding of all the critical factors for predicting first attempt failure.

Meta-Analysis

A total of 170 risk factors were extracted from the 20 included studies. Synonymous risk factors were combined where possible, and after consultation with subject matter experts, a total of 82 distinct risk factors were identified (see Appendix 2; http://links.lww.com/JIN/A111). Of those 82 distinct factors, only 22 were reported in at least 2 studies and, as such, were available for synthesis (Figure 3). History of DIVA was the most frequently reported risk factor, appearing in 13 of the 20 studies. Obesity was reported in 11 studies. Sex, vein palpability, and vein visibility were the next most common factors, each measured in 10 studies. MVMA-MF models on the entire set of 22 factors did not converge, likely due to the sparseness of data. Thus, 10 risk factors (history of DIVA, female sex, obesity, nonpalpable veins, nonvisible veins, chemotherapy, diabetes mellitus, drug abuse, age, and frequent PIVCs), which appeared in at least 4 studies, were included in the primary analysis (Figure 3). Only 1 study measured all 10 primary analysis risk factors.23

Figure 3.

Twenty-two risk factors were reported in at least 2 of the included studies after synonymous terms were combined. Risk factors reported in 2 or 3 studies are shaded in grey. Risk factors reported in 4 or more studies are shaded in black. Abbreviations: CAD, coronary artery disease; DIVA, difficult intravascular access; IV, intravenous; PVD, peripheral vascular disease.

The MVMA-MF analyses on the set of 10 risk factors measured in at least 4 studies found 5 factors that were significantly associated with DIVA (history of DIVA, obesity, non-palpable veins, non-visible veins, and IV drug abuse; Table 2). The primary analysis found 4 of the 6 clinician-identified set of a priori factors expected to be important in predicting DIVA to be significant (lack of vein palpability, obesity, lack of vein visibility, history of DIVA). Two of the 6 clinician-identified variables, history of dialysis and insertion site availability, were collected in too few studies to be sufficiently powered. The only variable not identified a priori that was identified through the MVMF-MA analysis as a significant predictor of DIVA was IV drug abuse.

TABLE 2. Risk Factors That Were Reported in 4 or More Studies Were Assessed as Risk Factors for DIVA Using MVMA-MF Analysis.

Variable	OR (95% CrI)
Nonpalpable veins	9.13* (3.97, 16.30)
Nonvisible veins	7.05* (2.48, 16.03)
History of DIVA	6.25* (3.32, 10.12)
IV drug abuse	3.70* (2.06, 6.71)
Frequent IV	2.26 (0.51, 11.27)
Obesity	1.70* (1.38, 2.16)
Chemotherapy	1.59 (0.80, 2.95)
Diabetes	1.47 (0.98, 2.60)
Female	1.17 (0.90, 1.88)
Age (>65 y)	1.15 (0.97, 1.45)

Odds ratios for significant risk factors are indicated with an asterisk (*).

Abbreviations: CrI, credible interval; DIVA, difficult intravascular access; IV, intravenous; MVMA-MF, multivariate meta-analysis of multiple factors; OR, odds ratio.

As a sensitivity analysis, a UVMA was run on all 22 risk factors identified in 2 or more studies. Effect estimates were generally similar between the UVMA and the MVMA-MF, although the MVMA-MF had wider intervals of uncertainty (see Appendix 3; http://links.lww.com/JIN/A112).

For 13 of the 20 studies included in the analysis, DIVA was defined as a PIVC insertion requiring more than 1 attempt. A sensitivity analysis was conducted by repeating the primary analysis with only these 13 studies. History of DIVA, vein palpability, and vein visibility remained as significant factors. Obesity, which was identified as significant in the primary analysis, became nonsignificant, and drug abuse was not included due to inadequate reporting among the 13 studies (Table 3).

TABLE 3. Sensitivity Analysis of Risk Factors for DIVA with Only Studies That Defined DIVA as Requiring More Than 1 Attempt.

Variable	Primary Analysis OR (95% CrI)	Restricted DIVA Definition OR (95% CrI)
Nonpalpable veins	9.13* (3.97, 16.30)	10.33* (5.01, 19.26)
Nonvisible veins	7.05* (2.48, 16.03)	7.99* (2.74, 20.18)
History of DIVA	6.25* (3.32, 10.12)	4.79* (2.56, 7.57)
IV drug abuse	3.70* (2.06, 6.71)	Not available for inclusion
Frequent IV	2.26 (0.51, 11.27)	Not available for inclusion
Obesity	1.70* (1.38, 2.16)	1.55 (0.78, 6.55)
Chemotherapy	1.59 (0.80, 2.95)	1.16 (0.44, 5.76)
Diabetes	1.47 (0.98, 2.60)	Not available for inclusion
Female	1.17 (0.90, 1.88)	1.18 (0.56, 3.07)
Age (65+)	1.15 (0.97, 1.45)	Not available for inclusion

Significant risk factors are indicated with an asterisk (*).

Abbreviations: CrI, credible interval; DIVA, difficult intravascular access; IV, intravenous; OR, odds ratio.

Using the 5 variables the primary analysis identified as significant (vein palpability, vein visibility, history of DIVA, history of IV drug abuse, and obesity) an acronym-based mnemonic guideline was created for clinical use (SAFE): See, Ask, Feel, Evaluate BMI (Figure 4). The SAFE rule incorporates the best available evidence about risk factors for DIVA and is intended to encourage the recognition of DIVA prior to the first PIVC insertion attempt.

Figure 4.

The SAFE rule emblem combines the key risk factors identified in the MVMF-MA. Abbreviations: BMI: body mass index; DIVA, difficult intravascular access; IV, intravenous.

All 5 of the SAFE rule risk factors were identified as significant through the primary MVMF-MA analysis and are also present in at least a quarter of other published DIVA assessment instruments identified in a recent SLR.8 Some of the other factors present in those external prediction instruments were unable to be included in the current analysis due to lack of sufficient data, but those that were included were not found to be significant predictors of DIVA. The risk factors in the SAFE rule, although not yet weighted or clinically validated, are generally similar to those identified by other DIVA prediction tools. Moreover, the assessment instruments that have ORs available are similar to those found in the current analyses (Table 4).1,22,28,32

TABLE 4. Odds Ratios for Risk Factors in Recently Published DIVA Prediction Tools Compared to the Risk Factors Included in the SAFE Rule.

Included Risk Factors	A-DIVA28	Modified A-DIVA1	EA-DIVA22 Development	EA-DIVA22 Validation	A-DICAVE32	SAFE
No visible veins	3.63*	5.9*	Combined to vein evaluation		6.1*	7.05*
No palpable veins	4.94*	4.8*	Combined to vein evaluation		13.3*	9.13*
History of DIVA	3.86*	2.7*	15.1	10.9	7.7*	6.25*
IV drug abuse			Combined to vascular depletion			3.70*
Vascular depletion			4.1	4.3
Obese						1.70*
Overweight			1.2	0.6
Coagulative disorder			1.7	1.6
Neurovascular disease			2.5	2.7
Clinical examination of skin			1.6	1.2
Vein evaluation			4.8*	10.5*
Only one side available			1.6	1.4
Practitioner expectation of DIVA		2.6*
Vein diameter <3 mm after tourniquet	3.37*	3.5*
Unplanned indication for surgery	4.86*

Significant risk factors are indicated with an asterisk (*). Risk factors present in the SAFE rule are bolded.

Abbreviations: A-DICAVE, adult-difficult venous catheterization scale; BMI, body mass index; DIVA, difficult intravascular access; EA-DIVA, enhanced adult DIVA; IV, intravenous; SAFE, See, Ask, Feel, Evaluate BMI.

DISCUSSION

Several recent SLRs have highlighted the importance of patient and clinical burden associated with failed PIVC insertion and have demonstrated that various DIVA prediction tools and risk factor studies exist.8–10 Despite the abundant evidence, no tool appears to have been widely adopted in clinical practice, and patients with DIVA continue to suffer as a result.8 Although there have been a wide variety of potential DIVA indicators identified throughout the risk factor and prediction tool literature, there are a few key factors that are consistently recognized. In a 2022 SLR of DIVA assessment of instruments, clinical practice guidelines, and escalation pathways, Paterson et al8 found that 88% of the tools they identified considered vein invisibility with or without a tourniquet to be an indicator of DIVA, and 69% included impalpable veins with or without a tourniquet. History of DIVA, vascular depletion (including damage from IV drug abuse), and overweight or obesity were reported in 38%, 31%, and 25% of assessment instruments, respectively.8 While a handful of these tools have been clinically validated, most were developed and tested for specific populations or clinical settings, and none incorporated the breadth of available data to inform the DIVA prediction risk factors they included.1,29,33 The present SLR and meta-analysis combined data from existing DIVA prediction tools and risk factor studies, along with clinician gestalt, to determine the best available indicators for predicting difficult venous access. The findings were synthesized into the SAFE rule, an evidence-based mnemonic guideline that stands for See, Ask, Feel, Evaluate BMI and aims to bring the intended prospective identification of DIVA from existing prediction tools into a format that can be easily and consistently implemented in clinical practice.

The first adult DIVA (A-DIVA) prediction tool was developed for surgical patients and was originally published in 2016 and validated in 2019 by van Loon et al.1,29 The A-DIVA was modified into the enhanced adult DIVA scale (EA-DIVA), which was developed and validated by Civetta et al in 2019.17 Next, the adult difficult venous catherization scale (A-DICAVE) was published in 2020 by Salleras-Duran et al33 for use in the emergency department. While all these DIVA prediction tools agree in their designation of certain risk factors as key, they also share an important fault: complexity of implementation. The tools have scoring or scaling systems that require a level of careful consideration that may not always be feasible to provide, especially when standards of care do not emphasize first-attempt insertion success. Although the phenomenon has not been formally reported, it is suspected that the existing DIVA assessment tools have not been widely adopted because their implementation may be viewed as inconvenient and an inefficient use of time that could be spent making additional insertion attempts. The SAFE rule may be faster, easier to implement, and could help guide decision-making in situations where DIVA may be present but time constraints prevent the use of a more complex tool. Consideration of the SAFE rule before every catheterization attempt could be a step toward reducing the unnecessary burdens associated with DIVA. It should also be noted that some risk factors, such as vein visibility and palpability, are encounter-specific and may change considerably, depending on the health status of the patient. Thus, although history of DIVA is another broadly acknowledged risk factor, a patient's venous access status has the potential to change and should, therefore, be assessed prior to each PIVC insertion encounter.

Multiple failed insertion attempts associated with DIVA are a threat to the efficient use of health care resources and lead to unnecessary burdens for patients, clinicians, and the broader health care system. When many insertion attempts are needed, there may be up to a 7-fold increase in costs, including those from additional supplies used and increased procedure times.6,8,34 Early escalation for patients with DIVA could save resources by avoiding the waste of time and supplies associated with multiple insertion attempts. Difficult-access patients who are treated with advanced techniques early on in PIVC placement may also be spared the pain and anxiety that can be associated with multiple venipunctures. When DIVA is anticipated or an insertion attempt has failed, escalation pathways and clinical practice guidelines unanimously recommend the use of vein visualization technology, and although these advanced technologies may be somewhat more resource intensive to use, it stands to reason that they could ultimately save costs for patients with DIVA and are likely to reduce stress and burdens for patients and clinicians alike.6,8

The present SLR identified 170 DIVA indicators, which were synthesized into 82 risk factors, after combining synonymous terms. Despite this seemingly large number of identified risk factors, there were several areas where data were lacking, and the analyses in the current study were less robust as a result. Clinical consultations pointed to dialysis as a risk factor likely to be important for predicting DIVA, but the data on dialysis as a risk factor were very limited. Smaller veins and a lack of available sites for cannulation were also less reported than anticipated. Another notable gap was the relative lack of data from people with dark skin. In the absence of sufficient data to analyze skin shade, race was considered in the current analysis but, of the 5 studies that did report race, only 3 considered it as a risk factor, so it could not be synthesized in the primary analysis. Given the prevalence of DIVA, there is a growing demand for standardization of reporting terminology around risk factors that would allow better research to be conducted in the future.10 Better data availability may impact the estimated importance of these and other sparsely reported factors.

Hospitalized patients with DIVA may be at a higher risk for catheter-related complications or failure and may require many reinsertions throughout their hospital stay.35 While the early detection of patients with DIVA may help to improve first-attempt success, it is important to note that these successes cannot be directly correlated with the catheter functionality and potential for subsequent failure. It is well documented that repeated failed insertion attempts can contribute to venous depletion, but it does not necessarily follow that a successful first attempt would protect against the conditions that would necessitate eventual reinsertions.10 Factors such as site and vein selection, catheter size, and method of securement may influence the longevity of the catheter and play a role in the requirement for reinsertions.35 Additional research is needed to describe the relationship between improved first-attempt success rates and catheter function and survival in patients with DIVA.

CONCLUSION

Difficult intravascular access is a common problem that is associated with a frequently performed procedure for hospitalized patients. Despite the prevalence of this problem and the breadth of literature describing risk factors associated with DIVA and encouraging improvements to first attempt success rates, no tool has been broadly implemented into clinical practice, and many standards of care continue to allow multiple attempts to place a PIVC. The current study is part of a growing body of evidence that suggests that a change in the standard of care is needed and that health care providers should aim to assess risk of DIVA before the first insertion attempt, with the goal of improving first-attempt success rates. The SAFE rule presents a viable, although not-yet clinically validated, tool for practical use in anticipating DIVA and improving first-attempt success rates. Given the extensive evidence to support the prospective identification of DIVA, the authors recommend an update to the recently proposed DIVA definition, “when a clinician has two or more failed attempts at PIV access using traditional techniques, physical examination findings are suggestive of DIVA (eg, no visible or palpable veins) or the patient has a stated or documented history of DIVA,” to allow for advanced recognition of patients with DIVA. Hence, the authors propose a modification to the definition, as follows: “Difficult Intravascular Access (DIVA): Refers to any patient, prior to any insertion attempt, that physical assessment yields no visible or palpable vasculature, the patient has a stated or a documented history of difficulty with obtaining vascular access, IV drug abuse, or has a BMI greater than 30. These patients should be classified as a DIVA patient and should be considered for escalation of advanced insertion techniques such as the use of ultrasound.”

Supplementary Material

jinfn-47-096-s001.docx ^{(81KB, docx)}

jinfn-47-096-s002.docx ^{(74KB, docx)}

jinfn-47-096-s003.docx ^{(78.1KB, docx)}