- Email: [email protected]
Evaluation of the performance of two different infusion pump systems in a neonatal intensive care unit
©MidwiferYLongman(21UK25Ltd 1989)G1989 roupS'
Midwffery
o,,oo,,8,.o,o~5oo21,,,o.oo
Evaluation of t h e p e r f o r m a n c e of t w o d i f f e r e n t infusion p u m p s y s t e m s in a neonatal intensive care unit Loredana Favetta, Mirella Lorenzutti, Daniela Petrina and Marina Cuttini
The aim of this study was to evaluate the performance of two different types of intravenous infusion pumps currently used in the Neonatal Intensive Care Unit of the Burlo Garofolo Children's Hospital, Trieste, Italy. During a 1-month-period two volumetric (Oximetrix Shaw) and four peristaltic (Ivac) pumps were studied. The total number of days of observation (pump-days) were 63 and 58 respectively. The accuracy, which is the difference between the volume of fluid to be infused according to the doctor's prescription and that actually delivered to the baby, expressed as a percentage on the daily programme, is satisfactory for both types of instruments, even if below the standard set by the manufacturers. Efficiency is evaluated through a separate set of criteria developed 'a priori' within each category of pumps, taking into account accuracy and the number of adjustments of infusion rate required per day. It appears that both volumetric and peristaltic pumps show in most days of observation a level of efficiency defined as good or medium: 81% and 72% respectively. For volumetric pumps this satisfactory performance is equally divided into the classes good and medium, whilst for peristaltic ones there is a predominance of medium days.
INTRODUCTION In the Neonatal Intensive Care Unit (NICU) of the Istituto per l'Infanzia Burlo Garofolo of Trieste, Italy, total parenteral nutrition for very Loredana Favetta, nurse Mirella Lorenzutti, nurse Daniela Petrina, nurse Marina Cuttini, MD, MPH Neonatologia Istituto per I'lnfanzia "Burlo Garofolo" Via Istria 65/1,34100, Trieste, Italy (Requests for offprints to MC)
Manuscript accepted 13 September 1988
low birthweight b a b i e s has been used with increasing frequency during the last years (Colonna et al, 1985). To minimise metabolic complications such as hyperglycaemia or to avoid such accidents as fluid overload in tiny patients (Colonna et al, 1986), it is essential to have intravenous infusion instruments which are precise, stable and reliable both for administration of small volumes of fluids, such as blood or drugs, and for large amounts of nutritional solutions. The purpose of this study was to assess the performance of two different types of intravenous in91
22
MIDWIFERY
fusion pump systems: the Oximetrix Shaw T M (model II/D) and the Ivac 531.
METHODS The Ivac 531 is an electronic peristaltic infusion pump (User Information, 1975) which is used in our Unit for the administration of fluid volumes less than 300 millilitres (ml) per day. The delivery rate is set at a number of drops per minute (dr/min) and is controlled by a photoelectric cell. A standard intravenous (iv) administration set is used with a normal or with a micro drop device. The alarm system is activated when the bottle is empty, when infusion is hindered, or because of power failure. The Oximetrix Shaw is a volumetric electromechanical pump (Oximetrix Service Manual, 1980) used in our Unit for the administration of large amounts of fluid: over 300 ml per day. The infusion rate is programmed at a number of millilitres per hour (ml/h). A 'dose-limit' function allows for preselection and delivery of a limited volume (dose) of fluid and signals when the dose is completed. Cumulative fluid volume delivered is indicated by a volume display, and retained in electronic memory until cleared. A disposable Shaw Pump Set with in-line pump cassette and air trap chamber is used. Alarm indicators signal an empty bottle, flow rate error, dose complete and low battery conditions. This study was performed during a one month period on five babies in the N I C U receiving iv nutrition through a deep vein 24 hours a day. Four Ivac and two volumetric pumps were used, and all of them were working for 24 hours per day. Two pumps (one peristaltic and one volumetric)were replaced for routine maintenance. The nutritional solutions were prepared daily by the hospital pharmaceutical service according to a physician written prescription. Administration sets were prepared following the instructions given by the manufacturers. The volume of fluid to be administered over the 24 hours was calculated according to the doctor's order (daily programme) as micro dr/min for Ivac or ml/h for volumetric pumps.
The infusion rate was checked by nurses about every 3 hours by reading the volume infused on the bottle scale of the Ivac pumps and on the cumulative volume display of the volumetric ones. If the rate was too slow or too fast according to the schedule, necessary corrections were carried out. At the end of the day the residue of fluid in the bottle was measured by means of a syringe and therefore the precise amount delivered to the baby was calculated. During the period of the study we recorded the following: 1. Discrepancy between the volume of fluid to be infused over the 24 hours according to physician's prescription and the volume really administered to the baby ('true difference'). 2. Number of times the nurses checked the infusion rate during the 24 hours. 3. Number of adjustments of the infusion rate made during 24 hours by the nurses to conform to the programme. 4. Maximum change of infusion rate, expressed in ml/h or equivalently in micro dr/min according to pump used, in comparison to the initial selected rate. For the Oximetrix pumps we also recorded the following items: 1. Discrepancy between the dose administered to the baby as signalled by the pump volume display, and that assessed by the nurse through manual measurement of the residue. 2. Discrepancy between the administered dose as signalled by the pump volume display and that programmed by the doctor's written order for tile 24 hours. In our Unit, as stated above, the two types of pumps are used in a different way and hence they are not directly comparable. Therefore, a set of criteria was designed separately for each of the two categories of pumps, in order to classify the performance of each single instrument as good, medium or bad (Table 1). The criteria take into account the difference between the volume of fluid really delivered to the baby and that prescribed by the physician, (true difference) expressed as a percentage on a daily programme, together with the number of adjustments of the
MIDWIFERy
23
Table 1 Performance classification criteria Criteria Peristaltic pumps
Performance
Volumetric pumps
Good
True difference and number of adjustments True difference and number of adjustments True difference and/or number of adjustments
Medium Bad
infusion rate over the 24 hours. The last item is valued more loosely for Ivac pumps to allow for the larger number of times they are habitually checked. Data were subsequently analysed by the Chi square test for contingency tables. The unit of analysis was represented by a p u m p working for a 24 hours period ('pump-day').
FINDINGS The study was carried out on six intravenous infusion instruments: two Oximetrix and four Ivac pumps. The total number of days in which the pumps worked ('pump-days') were 63 and 58 respectively (Table 2). Table 2 Number of days of w o r k ('pump-days') by type of instrument Days of work ('pump-days')
Total
33 8
63 58
Type of instrument Volumetric Peristaltic
30 18
31
1
Table 3 shows how the two types of pumps are used differently in our unit. Volumetric instruments are used for the administration of volumes larger than for the Ivac: mean values 379.5 ml versus 182.9 ml per day. As the Oximetrix are in some way considered safer by nurses, they are controlled less often (five times compared to about eight times per day for the Ivac).
~<3% ~<1 ~<10% ~<3 > 10% /> 4
True difference and number of adjustments True difference and number of adjustments True difference and/or number of adjustments
~<3% ~<3 ~<10% ~<5 > 10% >5
Infusion rate adjustments are about three times less frequent than for peristaltic pumps and controlling for t h e number of Checks does not substantially modify this finding. On average the maximum amount of change of infusion rate is smaller for the volumetric than for the peristaltic pumps: 1.5 ml versus 3.8 ml per hour. The disparity between the amount of fluid prescribed by the doctor's order and that really administered to the baby (true difference) seems to be double for the volumetric than for the Ivac pumps. However if expressed as a percentage on total daily p r o g r a m m e (accuracy), it is about the same for each of them, 60/0. This value is, however, larger than that allowed by the manufacturers: 2% for Ivac and 1% for Oximetrix. For the volumetric pumps there is also a difference between the volume delivered to the baby according to instrument volume display and that according to the nurse's manual measurement, with an average of more than 25 ml per day. In Table 4 pumps working days are classified in three groups on the basis of the criteria previously defined. It appears that the performance of both volumetric and peristaltic infusion pumps was on most days of observation good or medium: 81% and 72% respectively. For volumetric pumps t h i s satisfactory performance is equally divided into the two classes good and medium, whilst for peristaltic ones there is a predominance of medium days (52%). The overall Chi square (2 df) for difference in performance between the two types of instruments is 5.99 (p = 0.05). The frequency of good working days is higher among the Oximetrix (41% versus 21%) and the difference is statistically significant (Chi square 1 df 5.94, p <0.025).
24
MIDWIFERY
Table 3 Patterns of use and accuracy by type of instrument Volumetric pumps
Peristaltic pumps
Mean
Range
Mean•
Range
200-580 0-110 0-36 0-8 0-3
182.9 11.2 6.1 7.9 2.9 0.4 3,8
100-350 0-70 0-28 O-8 0-7
Daily program (ml/day) 379.5 True difference (ml) 20.2 True difference (%) 5.9 Number of checks/24 hour 5 Number of adjustments/24 hour 0.9 Number of adjustments/number of checks 0.2 Maximum change of infusion rate/24 h (ml/h) 1.5
The frequency of bad functioning is 19% for Oximetrix and 28% for Ivac pumps (Chi square 1 df 1.24, p>0.25).
DISCUSSION The findings of this study show that both volumetric and peristaltic pumps perform with a degree of accuracy which may be considered satisfactory, even if inferior to that described by the manufacturers. Volumetric instruments seem to be working with more efficiency according to our criteria, more definitely good and fewer bad pump-days. These findings are consistent with the feeling of nurses which from the beginning appeared to trust more the volumetric rather than the peristaltic pumps. It is important to recognise that the two types of instruments were never directly compared to each other, but only within themselves or to the standard set by the producers. This avoids the possible biases due to different model features, pattern of use, and even nurses' expectations. Table 4 Pumps performance according to criteria Volumetric pumps
No.
(%)
Peristaltic pumps
No.
(%)
Good Medium Bad
26 25 12
(41.3) (39.7) (19.0)
12 30 16
(20.7) (51.7) (27.6)
Total
63
(100.0)
58
(100.0)
0-8
0-9
The system we used to classify efficiency is an arbitrary choice. The criteria were defined by nurses 'a priori', that is before seeing the data, on the basis of their clinical experience, and exactly reflect their judgement about pump performance during a day of work. Sensitivity analysis carried out over a range of criteria variations did not substantially modify the findings. Sources of error other than pump malfunctioning can be hypothesised, such as mistakes made by the pharmacist in preparing the intravenous solutions, or by the nurses themselves in calculating the infusion rate or velocity adjustments. Random measurements of the volumes of nutritional solutions prepared by the pharmaceutic service were carried out to evaluate their accuracy in conforming to the doctor's prescription and a high degree of concordance was found. Nurses' mistakes are assumed to be randomly distributed between the two categories of pumps. Besides, as our study is an 'in vivo' assessment of instrument performance carried out directly on babies, it is of course sensitive to environmental conditions. Personnel behaviour, and therefore even possible mistakes, are part of the evaluation. Post Script: IVAC have recently produced a new model of pump (560). However, the 531 will continue to be used in neonatal units for some time.
Acknowledgement The authors wish to thank Mrs C. Lewisfor her assistance in reviewing the English.
MIDWIFERY
References Colonna F, Candusso M, Davi F et al 1985 La nutrizioue parenterale centrale (NPC) nei neonati di peso inferiore a 1500 grarnmi. Problemi tecnici, organizzativi a metodologici. Minerva Pediatrica 37:645-652 Colonna F, Candusso M, Davi F et al 1986 La nutrizione
25
parenterale centrale (NPC) nei neonati di peso inferiore a 1500 grarnnrni. Problemi metabolici e nutrizionali. Minerva Pediatrica 38:83-94 User Information 1975 Ivac 530/531 Infusion Pump Service Manual 1980 Oximetrix Shaw Infusion Pump System, Model II/D
Midwffery
o,,oo,,8,.o,o~5oo21,,,o.oo
Evaluation of t h e p e r f o r m a n c e of t w o d i f f e r e n t infusion p u m p s y s t e m s in a neonatal intensive care unit Loredana Favetta, Mirella Lorenzutti, Daniela Petrina and Marina Cuttini
The aim of this study was to evaluate the performance of two different types of intravenous infusion pumps currently used in the Neonatal Intensive Care Unit of the Burlo Garofolo Children's Hospital, Trieste, Italy. During a 1-month-period two volumetric (Oximetrix Shaw) and four peristaltic (Ivac) pumps were studied. The total number of days of observation (pump-days) were 63 and 58 respectively. The accuracy, which is the difference between the volume of fluid to be infused according to the doctor's prescription and that actually delivered to the baby, expressed as a percentage on the daily programme, is satisfactory for both types of instruments, even if below the standard set by the manufacturers. Efficiency is evaluated through a separate set of criteria developed 'a priori' within each category of pumps, taking into account accuracy and the number of adjustments of infusion rate required per day. It appears that both volumetric and peristaltic pumps show in most days of observation a level of efficiency defined as good or medium: 81% and 72% respectively. For volumetric pumps this satisfactory performance is equally divided into the classes good and medium, whilst for peristaltic ones there is a predominance of medium days.
INTRODUCTION In the Neonatal Intensive Care Unit (NICU) of the Istituto per l'Infanzia Burlo Garofolo of Trieste, Italy, total parenteral nutrition for very Loredana Favetta, nurse Mirella Lorenzutti, nurse Daniela Petrina, nurse Marina Cuttini, MD, MPH Neonatologia Istituto per I'lnfanzia "Burlo Garofolo" Via Istria 65/1,34100, Trieste, Italy (Requests for offprints to MC)
Manuscript accepted 13 September 1988
low birthweight b a b i e s has been used with increasing frequency during the last years (Colonna et al, 1985). To minimise metabolic complications such as hyperglycaemia or to avoid such accidents as fluid overload in tiny patients (Colonna et al, 1986), it is essential to have intravenous infusion instruments which are precise, stable and reliable both for administration of small volumes of fluids, such as blood or drugs, and for large amounts of nutritional solutions. The purpose of this study was to assess the performance of two different types of intravenous in91
22
MIDWIFERY
fusion pump systems: the Oximetrix Shaw T M (model II/D) and the Ivac 531.
METHODS The Ivac 531 is an electronic peristaltic infusion pump (User Information, 1975) which is used in our Unit for the administration of fluid volumes less than 300 millilitres (ml) per day. The delivery rate is set at a number of drops per minute (dr/min) and is controlled by a photoelectric cell. A standard intravenous (iv) administration set is used with a normal or with a micro drop device. The alarm system is activated when the bottle is empty, when infusion is hindered, or because of power failure. The Oximetrix Shaw is a volumetric electromechanical pump (Oximetrix Service Manual, 1980) used in our Unit for the administration of large amounts of fluid: over 300 ml per day. The infusion rate is programmed at a number of millilitres per hour (ml/h). A 'dose-limit' function allows for preselection and delivery of a limited volume (dose) of fluid and signals when the dose is completed. Cumulative fluid volume delivered is indicated by a volume display, and retained in electronic memory until cleared. A disposable Shaw Pump Set with in-line pump cassette and air trap chamber is used. Alarm indicators signal an empty bottle, flow rate error, dose complete and low battery conditions. This study was performed during a one month period on five babies in the N I C U receiving iv nutrition through a deep vein 24 hours a day. Four Ivac and two volumetric pumps were used, and all of them were working for 24 hours per day. Two pumps (one peristaltic and one volumetric)were replaced for routine maintenance. The nutritional solutions were prepared daily by the hospital pharmaceutical service according to a physician written prescription. Administration sets were prepared following the instructions given by the manufacturers. The volume of fluid to be administered over the 24 hours was calculated according to the doctor's order (daily programme) as micro dr/min for Ivac or ml/h for volumetric pumps.
The infusion rate was checked by nurses about every 3 hours by reading the volume infused on the bottle scale of the Ivac pumps and on the cumulative volume display of the volumetric ones. If the rate was too slow or too fast according to the schedule, necessary corrections were carried out. At the end of the day the residue of fluid in the bottle was measured by means of a syringe and therefore the precise amount delivered to the baby was calculated. During the period of the study we recorded the following: 1. Discrepancy between the volume of fluid to be infused over the 24 hours according to physician's prescription and the volume really administered to the baby ('true difference'). 2. Number of times the nurses checked the infusion rate during the 24 hours. 3. Number of adjustments of the infusion rate made during 24 hours by the nurses to conform to the programme. 4. Maximum change of infusion rate, expressed in ml/h or equivalently in micro dr/min according to pump used, in comparison to the initial selected rate. For the Oximetrix pumps we also recorded the following items: 1. Discrepancy between the dose administered to the baby as signalled by the pump volume display, and that assessed by the nurse through manual measurement of the residue. 2. Discrepancy between the administered dose as signalled by the pump volume display and that programmed by the doctor's written order for tile 24 hours. In our Unit, as stated above, the two types of pumps are used in a different way and hence they are not directly comparable. Therefore, a set of criteria was designed separately for each of the two categories of pumps, in order to classify the performance of each single instrument as good, medium or bad (Table 1). The criteria take into account the difference between the volume of fluid really delivered to the baby and that prescribed by the physician, (true difference) expressed as a percentage on a daily programme, together with the number of adjustments of the
MIDWIFERy
23
Table 1 Performance classification criteria Criteria Peristaltic pumps
Performance
Volumetric pumps
Good
True difference and number of adjustments True difference and number of adjustments True difference and/or number of adjustments
Medium Bad
infusion rate over the 24 hours. The last item is valued more loosely for Ivac pumps to allow for the larger number of times they are habitually checked. Data were subsequently analysed by the Chi square test for contingency tables. The unit of analysis was represented by a p u m p working for a 24 hours period ('pump-day').
FINDINGS The study was carried out on six intravenous infusion instruments: two Oximetrix and four Ivac pumps. The total number of days in which the pumps worked ('pump-days') were 63 and 58 respectively (Table 2). Table 2 Number of days of w o r k ('pump-days') by type of instrument Days of work ('pump-days')
Total
33 8
63 58
Type of instrument Volumetric Peristaltic
30 18
31
1
Table 3 shows how the two types of pumps are used differently in our unit. Volumetric instruments are used for the administration of volumes larger than for the Ivac: mean values 379.5 ml versus 182.9 ml per day. As the Oximetrix are in some way considered safer by nurses, they are controlled less often (five times compared to about eight times per day for the Ivac).
~<3% ~<1 ~<10% ~<3 > 10% /> 4
True difference and number of adjustments True difference and number of adjustments True difference and/or number of adjustments
~<3% ~<3 ~<10% ~<5 > 10% >5
Infusion rate adjustments are about three times less frequent than for peristaltic pumps and controlling for t h e number of Checks does not substantially modify this finding. On average the maximum amount of change of infusion rate is smaller for the volumetric than for the peristaltic pumps: 1.5 ml versus 3.8 ml per hour. The disparity between the amount of fluid prescribed by the doctor's order and that really administered to the baby (true difference) seems to be double for the volumetric than for the Ivac pumps. However if expressed as a percentage on total daily p r o g r a m m e (accuracy), it is about the same for each of them, 60/0. This value is, however, larger than that allowed by the manufacturers: 2% for Ivac and 1% for Oximetrix. For the volumetric pumps there is also a difference between the volume delivered to the baby according to instrument volume display and that according to the nurse's manual measurement, with an average of more than 25 ml per day. In Table 4 pumps working days are classified in three groups on the basis of the criteria previously defined. It appears that the performance of both volumetric and peristaltic infusion pumps was on most days of observation good or medium: 81% and 72% respectively. For volumetric pumps t h i s satisfactory performance is equally divided into the two classes good and medium, whilst for peristaltic ones there is a predominance of medium days (52%). The overall Chi square (2 df) for difference in performance between the two types of instruments is 5.99 (p = 0.05). The frequency of good working days is higher among the Oximetrix (41% versus 21%) and the difference is statistically significant (Chi square 1 df 5.94, p <0.025).
24
MIDWIFERY
Table 3 Patterns of use and accuracy by type of instrument Volumetric pumps
Peristaltic pumps
Mean
Range
Mean•
Range
200-580 0-110 0-36 0-8 0-3
182.9 11.2 6.1 7.9 2.9 0.4 3,8
100-350 0-70 0-28 O-8 0-7
Daily program (ml/day) 379.5 True difference (ml) 20.2 True difference (%) 5.9 Number of checks/24 hour 5 Number of adjustments/24 hour 0.9 Number of adjustments/number of checks 0.2 Maximum change of infusion rate/24 h (ml/h) 1.5
The frequency of bad functioning is 19% for Oximetrix and 28% for Ivac pumps (Chi square 1 df 1.24, p>0.25).
DISCUSSION The findings of this study show that both volumetric and peristaltic pumps perform with a degree of accuracy which may be considered satisfactory, even if inferior to that described by the manufacturers. Volumetric instruments seem to be working with more efficiency according to our criteria, more definitely good and fewer bad pump-days. These findings are consistent with the feeling of nurses which from the beginning appeared to trust more the volumetric rather than the peristaltic pumps. It is important to recognise that the two types of instruments were never directly compared to each other, but only within themselves or to the standard set by the producers. This avoids the possible biases due to different model features, pattern of use, and even nurses' expectations. Table 4 Pumps performance according to criteria Volumetric pumps
No.
(%)
Peristaltic pumps
No.
(%)
Good Medium Bad
26 25 12
(41.3) (39.7) (19.0)
12 30 16
(20.7) (51.7) (27.6)
Total
63
(100.0)
58
(100.0)
0-8
0-9
The system we used to classify efficiency is an arbitrary choice. The criteria were defined by nurses 'a priori', that is before seeing the data, on the basis of their clinical experience, and exactly reflect their judgement about pump performance during a day of work. Sensitivity analysis carried out over a range of criteria variations did not substantially modify the findings. Sources of error other than pump malfunctioning can be hypothesised, such as mistakes made by the pharmacist in preparing the intravenous solutions, or by the nurses themselves in calculating the infusion rate or velocity adjustments. Random measurements of the volumes of nutritional solutions prepared by the pharmaceutic service were carried out to evaluate their accuracy in conforming to the doctor's prescription and a high degree of concordance was found. Nurses' mistakes are assumed to be randomly distributed between the two categories of pumps. Besides, as our study is an 'in vivo' assessment of instrument performance carried out directly on babies, it is of course sensitive to environmental conditions. Personnel behaviour, and therefore even possible mistakes, are part of the evaluation. Post Script: IVAC have recently produced a new model of pump (560). However, the 531 will continue to be used in neonatal units for some time.
Acknowledgement The authors wish to thank Mrs C. Lewisfor her assistance in reviewing the English.
MIDWIFERY
References Colonna F, Candusso M, Davi F et al 1985 La nutrizioue parenterale centrale (NPC) nei neonati di peso inferiore a 1500 grarnmi. Problemi tecnici, organizzativi a metodologici. Minerva Pediatrica 37:645-652 Colonna F, Candusso M, Davi F et al 1986 La nutrizione
25
parenterale centrale (NPC) nei neonati di peso inferiore a 1500 grarnnrni. Problemi metabolici e nutrizionali. Minerva Pediatrica 38:83-94 User Information 1975 Ivac 530/531 Infusion Pump Service Manual 1980 Oximetrix Shaw Infusion Pump System, Model II/D