Article Text

Download PDFPDF

Risk factors for difficult peripheral venous cannulation in hospitalised patients. Protocol for a multicentre case–control study in 48 units of eight public hospitals in Spain
  1. Miguel Angel Rodriguez-Calero1,2,
  2. Ismael Fernandez-Fernandez3,
  3. Luis Javier Molero-Ballester3,
  4. Catalina Matamalas-Massanet3,
  5. Luis Moreno-Mejias3,
  6. Joan Ernest de Pedro-Gomez4,
  7. Ian Blanco-Mavillard1,
  8. Jose Miguel Morales-Asencio5
  1. 1 Quality, Teaching and Research Unit, Hospital Manacor, Manacor, Balearic Islands, Spain
  2. 2 Department of Nursing, University of the Balearic Islands, Palma, Balearic Islands, Spain
  3. 3 Hospitalisation area, Hospital Manacor, Manacor, Balearic Islands, Spain
  4. 4 Department of Nursing and Physiotherapy, University of the Balearic Islands, Palma, Balearic Islands, Spain
  5. 5 Department of Nursing, University of Málaga, Málaga, Spain
  1. Correspondence to Miguel Angel Rodriguez-Calero; mianrodriguez{at}


Introduction Patients with difficult venous access experience undesirable effects during healthcare, such as delayed diagnosis and initiation of treatment, stress and pain related to the technique and reduced satisfaction. This study aims to identify risk factors with which to model the appearance of difficulty in achieving peripheral venous puncture in hospital treatment.

Methods and analysis Case–control study. We will include adult patients requiring peripheral venous cannulation in eight public hospitals, excluding those in emergency situations and women in childbirth or during puerperium. The nurse who performs the technique will record in an anonymised register variables related to the intervention. Subsequently, a researcher will extract the health variables from the patient’s medical history. Patients who present one of the following conditions will be assigned to the case group: two or more failed punctures, need for puncture support, need for central access after failure to achieve peripheral access, or decision to reject the technique. The control group will be obtained from records of patients who do not meet the above conditions. It has been stated a minimum sample size of 2070 patients, 207 cases and 1863 controls.

A descriptive analysis will be made of the distribution of the phenomenon. The variables hypothesised to be risk factors for the appearance of difficult venous cannulation will be studied using a logistic regression model.

Ethics and dissemination The study was funded in January 2017 and obtained ethical approval from the Research Ethics Committee of the Balearic Islands. Informed consent will be obtained prior to data collection. Results will be published in a peer-reviewed scientific journal.

  • venous catheterisation, peripheral
  • risk factors
  • difficult venous access
  • nursing
  • ultrasound
  • intravenous therapy

This is an Open Access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See:

Statistics from

Strengths and limitations of this study

  • To our knowledge, no previous case–control studies have been conducted to identify the risk factors for difficult peripheral cannulation, or to describe this problem in different healthcare settings.

  • Cases and controls will be reported by clinicians using the same source and of recordings. Blinding study hypothesis and criteria for the assignment to each group will ensure a reliable comparability.

  • Profiles of patients at risk are needed to improve decision-making regarding cannulation routes and techniques, and to ensure the suitability and maintenance of different devices.


Peripheral venous catheters (PVC) are the most commonly used invasive devices in hospital care.1 Although the insertion of a PVC is usually a simple technique, difficulty can arise in this cannulation, requiring multiple punctures before the device is correctly situated. Multiple punctures provoke delays in care, in obtaining diagnosis or in initiating treatment.2–4 Furthermore, it generates stress, heightens perceptions of pain5 and reduces satisfaction, both among patients and among the professionals performing the technique.4 6

In addition, multiple punctures may be associated with a progressive deterioration of the vascular tree, termed ‘vascular exhaustion’, which makes vascular access even more difficult in successive contacts with the patient.7


Although difficult peripheral intravenous cannulation (DPIVC) occurs in 10%–24% of adults and in up to 37% of children who require a peripheral route during hospital treatment, in many respects it is still insufficiently studied.4 Although there is no consensus among researchers as to the necessary conditions for considering a case as ‘difficult’, DPIVC is generally understood as arising when two or more punctures are performed without success, or when puncture support methods are required, or when the impossibility of obtaining peripheral access means that a central venous catheter (CVC) must be inserted.8 Most current research in this area addresses the development of puncture support techniques,2 especially ultrasound, and few studies have analysed DPIVC as a health problem, or the factors that may promote its appearance.

DPIVC is associated with a greater need for CVC, and studies have shown that a high percentage of the latter catheters are inserted not because of the patient’s therapeutic needs but because it is impossible to use a peripheral access catheter.9 This circumstance heightens both the number and the severity of complications associated with catheter access, such as local infection, bacteraemia, thrombosis and pneumothorax. These, and other complications, are in turn associated with increased duration of hospital stay, greater morbimortality and higher costs.9–11

Significant health benefits could be achieved by avoiding potentially unnecessary central catheters.12 For example, regarding bacteraemia related to venous catheterisation, which is the principal and most severe complication in this respect, the incidence is significantly higher for central catheters; thus, bacteraemia affects 2.7 cases per 1000 days of central catheterisation, but only 1.1 cases per 1000 days of peripheral intravenous central catheter (PICC) and 0.5 cases per 1000 days for peripheral access.10 13 Indeed, venous catheter-related bacteraemia may be considered an independent cause of hospital morbidity and mortality, as each case generates an additional 10–20 days of hospital stay and increases costs by between $4000 and $56 000.14

In fact, in many cases, CVCs are inserted unnecessarily. Studies have reported a reduction of 80%–85% in the use of CVC in hospital patients with DPIVC when specific programmes were implemented.9 Similarly, Stokowski et al in 200911 observed a marked reduction in PICC-related complications (bacteraemia, thrombosis, obstruction and accidental withdrawal) following the provision of a training programme for nurses in the use of ultrasound techniques for venous cannulation. Implementation of this programme also reduced variability among other health professionals involved (radiologists, surgeons and anaesthetists), producing cost savings of $C270–$C305 for each catheter inserted. A similar programme, conducted in Texas, USA, achieved a 74% reduction in the number of CVCs inserted (including intensive care), mainly by replacing them with PICCs, which were inserted by nurses trained in the use of ultrasound techniques.15 This intervention reduced costs by US$200 000 per year, or US$1614 per PICC inserted.

Risk factors for DPIVC

It has been argued that strategies should be promoted to avoid multiple punctures and the undesirable effects of central access catheterisation.16 Although there is a growing body of evidence in favour of cannulation support methods (ultrasound, infrared and transillumination), few studies have attempted to identify risk factors for DPIVC or the profiles of patients likely to present it. To our knowledge, the only studies conducted in this area, to date, have been limited to specific hospital areas (intensive care, Accident and Emergency (A&E), paediatrics and oncology), and so there is little scope for comparing different approaches. Specifically, it has been suggested that several advanced chronic conditions may contribute to the progressive degradation of the peripheral vascular tree, such as obesity, vasculopathy and chronic pluripathology.7 17–20 However, these studies focus on the application of ultrasound to improve the effectiveness of puncture techniques, and so their approach to potential risk factors should be considered with caution.

In the context of hospital A&E services, three earlier studies have made interesting findings.

Sebbane et al conducted a study in France in 2013, without a control group, evaluating risk factors that determine the success of the first attempt at cannulation.21 These authors observed an association between extreme values for body mass index and the appearance of DPIVC, which was also associated with poor assessment by the health professional (whether doctor or nurse) of the viability of access. In fact, the professional’s view of the feasibility of cannulation has been explored in various studies, many of which have found it to be a relevant factor and a possible predictor of difficulty in obtaining venous access.22 Another study concluded that certain variables related to the professional who performs the technique, regarding his/her professional experience in general and concerning venous cannulation in particular, may also influence the effectiveness of the intervention.23

In 2016, Carr et al performed a cohort study which sought to identify factors relevant to the success of venous cannulation in patients treated at hospital A&E units.24 These authors, too, highlighted the importance of the professional’s assessment of the viability of venous access (visibility and palpability of the vein), in addition to factors such as cachexia (wasting syndrome) and advanced age, which were potentially associated with difficulty. This study also identified differences related to the location of the vein to be punctured and to the cannulation experience of the clinician performing the technique.

Finally, Fields et al reported that previous pathological conditions, such as diabetes, parenteral drug abuse and spindle cell disease, can increase the risk of DPIVC.25 Other relevant factors, although to a lesser extent, were previous episodes of puncture difficulty and the need for puncture support systems, observed in previous contacts with the patient.

In view of this background, we consider it necessary to analyse, in a single study, the different variables that have been proposed as potential risk factors for difficulty in cannulation, including care settings other than hospital A&E units.

In this project, we aim to identify the risk factors affecting patients with DPIVC, and to determine the weight of each of these factors, so that a model can be established by means of which patients at risk can be identified at an early stage and so that puncture support methods can be prioritised,26 taking into account that the use of such methods is increasingly recommended.27



The presence of potential risk factors considered in the study will independently increase the risk of the patient to present difficulty during peripheral venous cannulation.


The main study goal is to identify the possible risk factors associated with the patient, thus enabling us to establish a model with which to estimate the probability of difficult access to venous cannulation in hospital treatment.

Secondary goals:

  • to determine the characteristics of patients with DPIVC according to different care profiles (medical hospitalisation, surgical hospitalisation, surgical area and A&E intensive care);

  • to describe the type of venous catheter insertion technique according to the appearance of difficulty in cannulation: number of punctures required, perception of pain, resources needed (number of professionals and estimated time required) and need for alternative methods (CVC, ultrasound support, referral to other professionals);

  • to determine whether the occurrence of such difficulty is influenced by the experience and characteristics of the health professional involved.

Methods and analysis


Case–control study with incident cases.


Forty-eight units, corresponding to different care settings: A&E, intensive care, surgical area and hospitalisation units, in eight public hospitals in the Spanish National Health System, with diverse profiles, including three university hospitals and five second-level hospitals.


Adults (18 years old or more) for whom peripheral pathway cannulation is performed or attempted, and who consent to participate in the study. Patients in emergency situations and women during childbirth or puerperium will be excluded.

Data collection

The nurse who performs the technique will record, in an anonymised record, the variables related to the intervention and the patient’s medical history number. All nurses may add such records during the study period from 1 February to 31 December 2017. Members of the research team will retrospectively review the medical history to compile the health variables. The data collection system was piloted in four of the above hospital units in February 2016.

All patients in the sample population who present DPIVC at some time will be included in the case group if they meet at least one of the following conditions: two or more failed punctures; the need for puncture support techniques (ultrasound, infrared or transillumination) when accessible vessels cannot be identified (excluding ultrasound scans for other purposes); the need for central access after failure to achieve peripheral access or decision not to implement it (no venous access achieved and the procedure is abandoned). Subsequently, we will determine the distribution of the incidence of DPIVC by hospital environments (medical hospitalisation units, surgical hospitalisation units, surgical area and A&E intensive care). To offset the effects of possible differences in the inclusion of patients and their different profiles according to the units participating in the study, the control group will be selected by random sampling stratified by the same treatment environments, following the distribution of incidence observed in the case group. This sample will be composed of the patients included in the study who do not present the conditions for selection to the case group. Three controls will be selected for each case. The nurses will be blinded to the selection criteria for cases and controls to avoid selection bias.

Since the study will require the involvement of a significant number of professionals from different environments, a team of collaborators has been recruited, all of them registered nurses, to coordinate the study in their respective units and centres, thus serving as a bridge between the research team and the other professionals.

Variables and definitions

Taking into account previous studies in this field, 13 variables will be hypothesised as possible risk factors. Variables will also be considered to assess the comparability of the case and control groups. Table 1 lists these variables and their definitions.

Table 1

Variables and definitions


The minimum sample size was calculated taking as a reference the risk factor ‘diabetes’ from the study of A&E patients conducted by Fields et al in 2014.5 Assuming an alpha risk of 0.05 and a beta risk of 0.2 in a bilateral test, we calculated that 87 cases and 261 controls would be required to detect a minimum OR of 2.1, assuming a rate of exposure of 0.5 in the control group. In this consideration, the Poisson method was used. In addition, another 10 cases were attributed by category, following the system described by Peduzzi et al,28 and so the total minimum sample required is 207 cases. Assuming a frequency of 10%, 2070 patients must be identified to achieve the population size required for the case group. The estimated time to reach this sample size is 10 months, although this could be extended if necessary.

Data analysis

A descriptive analysis of the variables will be performed, including the distribution of the phenomenon by hospital environments and services (type of attention). Tests of association will be applied between the main study variables: hypothetical risk factors, characteristics of the technique, environment and experience of the professional. The association will be determined by bivariate analysis based on Χtest, Student’s t-test, Mann-Whitney U test, Wilcoxon W and Friedman tests, analysis of variance, and Pearson and Spearman correlations, depending on the nature and normality of distribution of the variables. Subsequently, the variables hypothesised as risk factors for the onset of DPIVC will be analysed using a logistic regression model to obtain the respective adjusted ORs.

Validity and reliability/rigour

The fact that cases and controls will be recruited from the same source, together with the inclusion of sample adjustment variables, will ensure the reliable comparability of the groups.

The control group will be distributed by stratified random sampling, which will ensure the homogeneity of the case and control groups.

Blinding to the study hypothesis and to the variables hypothesised as risk factors will prevent any selection bias that might arise in the nurses participating in the data collection process.

Relevant variables will be considered in order to study the possible influence of the professional profile of the nurse performing the technique on the appearance of DPIVC.

The multicentre nature of the study and the inclusion of different hospital profiles, and of hospitals located in different geographic areas, will enhance the diversity of the sample and its external validity.


The variable ‘spindle cell disease’ is not included in our study because of its low prevalence in the reference population. Since our study focuses on patient risk factors, variables related to the nurse’s experience have not been hypothesised as potential risk factors. An association analysis of these variables will be conducted to determine whether future studies in this regard are needed.

Ethics and dissemination

The study does not involve intervention or change in usual practice. The patients will be asked by the clinician nurse to give their signed informed consent, and will be provided with clearly written information about the purpose and implications of the research.

The computerised database does not contain patient identification. The individuals involved in compiling data will sign a confidentiality agreement.

The results of this study will be sent to a peer-reviewed scientific journal for publication.


The proposed study will enable us to obtain profiles of patients at risk of difficulty in peripheral venous cannulation. Identifying this risk at an early stage will facilitate the early and selective use of puncture support methods such as ultrasound or infrared imaging.

Nurse-led intravenous treatment teams can use this information to identify priority patients and to ensure the appropriateness of the interventions made. The information obtained regarding the use of nursing resources for managing DPIVC may also be useful for these teams.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.


  • Contributors MARC coordinates the research team and is responsible for reporting Ethics Committees and every institution involved in the study. MARC and IBM revised previous literature about puncture difficulty in adults, and designed data collection methods along with LJMB, IFF, CMM and LMM. JEPG and JMMA audited the study design, especially concerning statistical analysis proposed. IFF designed and maintains the database to ensure data validity. MARC, IBM, JEPG and JMMA wrote first version of the protocol, which was later edited by all the authors.

  • Funding This work was supported by the Balearic College of Nursing (grant number PI453/2016).

  • Competing interests None declared.

  • Patient consent Obtained.

  • Ethics approval Research Ethics Committee of the Balearic Islands (CEI_IB). Approval code IB3137/16PI.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.