Table 3 Characteristics of studies included in the review

[26]

N. Samaranayake et al., 2012, Hong Kong

1538 incident reports involving IP were randomly selected and evaluated. The pumps were of varying models and makes that were not specified.

Technology-related errors were analyzed to identify the type of technology responsible and the causes. Errors were divided into (i) Technology-related errors; (ii) Errors not related to technology. Within group (i) there was a division into: sociotechnical errors (which have human interaction) and device errors (related to technical defects of the device).

Of the errors involving technologies, 75.3% were related to prescription, 14.8% to medication administration, 8.4% to dispensing and 2.7% to other causes. Infusion pumps were responsible for 5.8% of sociotechnical errors. Device errors have only been observed with the use of infusion pumps.

[27]

A. Golpaygani et al., 2017, Iran

50 infusion pumps from 04 different brands and models were not specified.

Tests conducted to verify that the device pumps the required flow rate, volume and bolus with the required precision rate; verified that occlusion alarms are activated during emergency conditions and the device is safe for patient and operator use. In addition, electrical safety tests were performed.

The quantitative analysis of the variables related to the precision of the flow showed that the results obtained are critical and that the percentage of infusion error can be above 20% depending on the conditions of the equipment.

[28]

A. Cauchi et al., 2013, London

19 BBraun brand infusion pumps and Infusomat Space model.

Record errors in typing prescriptions were analyzed using the 5-key keyboard; the results were used to determine the probability distribution for the errors found.

For the set of infusion pumps studied, a percentage of typing errors of 7.13% was found for the infusion volume (VTBI) and 16.91% for the infusion rate or rhythm (rate).

[29]

S. Taghipour et al., 2011, Canada

Hospital history data from various pumps.

Different failure modes were considered, of different models of infusion pumps, being studied failures of the warning system (audible), chassis and battery.

The authors found very similar results in relation to the reference method, considering censored data.

[30]

K. Giuliano, 2018, United States

Volumetric infusion pump recall data.

Analyzes the number and type of failures in the different classes of infusion pumps.

The researchers point to the need to develop new technologies associated with the design of infusion pumps, aimed at improving usability and safety.

[25]

E. Batista et al., 2021, Portugal

Nexus 3000 infusion pump.

Calibration and uncertainty calculation; validation of the interferometry method, comparing with the gravimetric method.

Determination of uncertainty and validation of the proposed method.

[31]

M. Etelvino et al., 2019, Brazil

371 peristaltic volumetric infusion pumps from two different brands.

They carried out descriptive research using a qualitative approach in a hospital located in Rio de Janeiro. The analyzes were based on IP maintenance records and on collections made from April to June 2017.

As for the record for the last preventive maintenance performed, it was identified that less than 10% of the equipment studied were up to date, 54.5% had an expired preventive maintenance record, 5.9% had an illegible preventive maintenance record and 29.9% lack of preventive maintenance records.

[32]

M. Blancher et al., 2019, France

05 syringe infusion pumps, two of the standard model and three of low weight.

Benchtop comparative study with two different flow measurement methods performed during a 2-h infusion period at altitudes of 300, 1700 and 3000 m.

Lightweight models provide bolus rather than continuous flow. Even though they are 10x heavier, standard devices seem to be more reliable even at different altitudes.

[33]

S. Lee et al., 2018, Korea

Pressure-based pump, syringe pump, and drug delivery device. Syringe pump (Chemyx, Nexus 3000), commercial pump based on pressure (Elveflow).

They evaluated infusion pumps using the dynamic gravimetric method. Flow and pulsating flow stabilities from syringe and infusion pumps were analyzed according to flow rate. The measurement error and its uncertainty were obtained according to the flow.

The authors concluded that pulsating flows and rates can be measured as they interfere with pump operation.

[23]

E. Batista et al., 2019, Portugal

Syringe pumps and peristaltic pumps.

The study investigated the influence of rapidly changing flow rates due to a predefined flow rate change. In addition, a multi-infusion setup is developed to analyze fluid flow rates and their compositions at the exit of the infusion line.

By improving the precision of flow rate measurement of drug delivery devices, with the development of new measurement methods, dosing errors can be reduced. This can be achieved by capturing traceable calibrations of low-flow and ultra-low-flow infusion devices and by better understanding the calibration of dosing administration in clinical settings, especially in the cases of multiple infusion systems.

[34]

S. Manrique Rodríguez et al., 2014, Spain

Smart infusion pumps.

Analysis of failure modes and effects (FMEA) in the pediatric intensive care unit of a General and Teaching Hospital. The FMEA was carried out before the implementation of CareFusion smart infusion pumps and 18 months after identifying the risk points during three different stages of the implementation process: creation of a drug library; using the technology during clinical practice and analyzing the stored data using Guardrails R CQI v4.1 Event Reporter software.

Several improvement actions were carried out, including periodic reviews of the drug library, the development of supporting documents and profile training in the system. Eighteen months after implementation, these measures helped to reduce the likelihood of each risk point occurring and increase the likelihood of its detection.

[24]

E. Batista et al., 2020, Portugal

Syringe pump, brand not identified.

Investigation of different flow regimes, liquid mixing behavior and occlusion phenomena in infusion systems aimed at improve dosing accuracy mainly at flows as low as 100 nL/min.

The results showed that the errors obtained using PP syringes are considerably higher than those obtained with glass syringes, due to the compliance effect. This indicates that PP syringes should not be used in micro flow measurements or in flow generators employed in bio analytical, medical or microfluidic applications unless the entire setup (syringe and flow generator) can be calibrated/evaluated by a recognized laboratory and traceable.

[35]

M. Felipe et al., 2020, Brazil

03 syringe infusion pumps.

The syringe pumps were placed at the distal outlet level of the infusion line, 30 cm above and 30 cm below, to verify how variations in the height and density of the solution can influence the accuracy of the pumps.

The position of the syringe infusion pump can influence the amount of volume infused. The influence was more evident in the low infusion rate. These variants should be considered in intravenous syringe therapy, in infusion pumps in pediatric patients, to reduce medication errors triggered by changes in hydrostatic pressure and in the system.

[36]

M. Baeckert et al., 2020, Switzeland

07 sets of syringe infusion pumps.

The sets were evaluated in an in vitro study during start-up, vertical displacement maneuvers and occlusion of the infusion line at a defined flow rate of 1 ml h\(\hat{}\)-1. The measured data were used as input to a concentration simulation pharmacokinetic simulation plasma during a continuous neonatal epinephrine infusion pattern.

The problems that affected all tested sets are mainly related to the working principle of the infusion pump syringe and will only be partially resolved by incremental improvements to the existing equipment.

[37]

R. Gandillon 2013, United States

Data from 30000 general purpose infusion pumps from four manufacturers.

The intervention performed was found in a database to analyze the error rate as a function of cost and maintenance time.

By looking at historical repair information and choosing a provider with a low repair rate, many dollars and hours can be saved at a single hospital and, ideally, be reinvested in improving patient care. The study comes as a clinical utility in decisions about medical equipment acquisitions and fostering collaboration between clinical engineers, physicians and equipment manufacturers to improve medical devices, design and improve patient safety.