ONLINE FIRST OCTOBER 18, 2017 ORIGINAL RESEARCH

Transcription

1 ONLINE FIRST OCTOBER 18, 2017 ORIGINAL RESEARCH Oservtionl Study of Peripherl Intrvenous Ctheter Outcomes in Adult Hospitlized Ptients: A Multivrile Anlysis of Peripherl Intrvenous Ctheter Filure Nicole Mrsh RN, BN, MdvPrc 1,2,3 *, Jon Wester, RN, BA 1,2,4, Emily Lrsen, RN, BN, GDipHlthRes 1,2,3, Mrie Cooke, RN, PhD 2,3, Gor Mihl, Meng (Mech), GCert (Biostt) 2,3,5,6, Clire M. Rickrd, RN, PhD 1,2,3 1 Nursing & Midwifery Reserch Centre, Royl Brisne nd Women s Hospitl, Brisne, Queenslnd, Austrli; 2 School of Nursing nd Midwifery, Griffith University, Brisne, Queenslnd, Austrli; 3 Allince for Vsculr Access Teching nd Reserch Group, Menzies Helth Institute, Brisne, Queenslnd, Austrli; 4 School of Nursing, Queenslnd University of Technology, Brisne, Queenslnd, Austrli; 5 School of Medicine, Griffith University, Brisne, Queenslnd, Austrli; 6 Centre for Applied Helth Economics, Menzies Helth Institute, Brisne, Queenslnd, Austrli. BACKGROUND: Almost 70% of hospitlized ptients require peripherl intrvenous ctheter (PIV), yet up to 69% of PIVs fil prior to completion of therpy. OBJECTIVE: To identify risk fctors ssocited with PIV filure. DESIGN: A single center, prospective, cohort study. SETTING: Medicl nd surgicl wrds of tertiry hospitl locted in Queenslnd, Austrli. PARTICIPANTS: Adult ptients requiring PIV. MEASUREMENTS: Demogrphic, clinicl, nd potentil PIV risk fctors were collected. Filure occurred if the ctheter hd complictions t removl. RESULTS: We recruited 1000 ptients. Ctheter filure occurred in 512 (32%) of 1578 PIVs. Occlusion/infiltrtion risk fctors included intrvenous (IV) flucloxcillin (hzrd rtio [HR], 1.98; 95% confidence intervl [CI], ), 22-guge PIVs (HR, 1.43; 95% CI, ), nd femle ptients (HR, 1.48; 95% CI, ). Phleitis ws ssocited with femle ptients (HR, 1.81; 95% CI, ), ruised insertion sites (HR, 2.16; 95% CI, ), IV flucloxcillin (HR, 2.01; 95% CI, ), nd dominnt side insertion (HR, 1.39; 95% CI, ). Dislodgement risks were prmedic insertion (HR, 1.78; 95% CI, ). Ech increse y 1 in the verge numer of dily PIV ccesses ws ssocited (HR 1.11, 95% CI ) (HR 1.14, 95% CI ) with occlusion/infiltrtion, phleitis nd dislodgement. Additionl securement products were ssocited with less (HR 0.32, 95% CI ) (HR 0.63, 95% CI ) occlusion/infiltrtion, phleitis nd dislodgement. CONCLUSION: Modifile risk fctors should inform eduction nd inserter skill development to reduce the currently high rte of PIV filure. Journl of Hospitl Medicine 2017;12: XXX-XXX Society of Hospitl Medicine INTRODUCTION Peripherl intrvenous ctheter (PIV) insertion is the fstest, simplest, nd most cost-effective method to gin vsculr ccess, nd it is used for short-term intrvenous (IV) fluids, medictions, lood products, nd contrst medi. 1 It is the most common invsive device in hospitlized ptients, 2 with up to 70% of hospitl ptients receiving PIV. 3 Uncceptle PIV filure rtes hve een reported s high s 69%. 4-7 Filure is most frequently due to phleitis (vein wll irrittion/inflmmtion), occlusion (lockge), infiltrtion or extrvstion (IV fluids/vesicnt therpy entering surrounding tissue), prtil dislodgement or ccidentl removl, lekge, nd infection. 4,6,8 These filures hve importnt implictions for ptients, who endure the discomfort of PIV complictions nd ctheter replcements, nd helthcre stff nd udgets. *Address for correspondence nd reprint requests: Nicole Mrsh, RN, BN, MdvPrc, Center for Clinicl Nursing, Royl Brisne nd Women s Hospitl, Butterfield St, Herston, Queenslnd, 4029 Austrli; Telephone: ; Fx: ; E-mil: nicole.mrsh@helth.qld.gov.u Received: April 25, 2017; Revised: June 18, 2017; Accepted: June 21, Society of Hospitl Medicine DOI /jhm.2867 To reduce the incidence of ctheter filure nd void preventle PIV replcements, cler understnding of why ctheters fil is required. Previous reserch hs identified tht ctheter guge, 9-11 insertion site, nd inserter skill 10,15 hve n impct on PIV filure. Limittions of existing reserch re smll study sizes, retrospective design, 19 or secondry nlysis of n existing dt set; ll potentilly introduce smpling is. 10,20 To overcome these potentil ises, we developed dt collection instrument sed on the ctheter-ssocited risk fctors descried in the literture, 9-11,13 nd other potentil insertion nd mintennce risks for PIV filure (eg, multiple insertion ttempts, medictions dministered), with dt collected prospectively. The study im ws to improve ptient outcomes y identifying PIV insertion nd mintennce risk fctors menle to modifiction through eduction or lterntive clinicl interventions, such s ctheter guge selection or insertion site. METHODS Study Design nd Prticipnts We conducted this prospective cohort study in lrge tertiry hospitl in Queenslnd, Austrli. Ethics committee pprovl ws otined from the hospitl (HREC/14/ An Officil Puliction of the Society of Hospitl Medicine Journl of Hospitl Medicine Pulished Online Octoer 2017 E1

2 Mrsh et l Peripherl Intrvenous Ctheter Outcomes QRBW/76) nd Griffith University (NRS/26/14/HREC). The study ws registered with the Austrlin New Zelnd Clinicl Trils Registry (ACTRN ). Ptients in medicl nd surgicl wrds were screened Mondy, Wednesdy, nd Fridy etween Octoer 2014 nd Decemer Ptients over 18 yers with PIV (BD Insyte TM Autogurd TM BC; Becton Dickinson, Frnklin Lkes, NJ) inserted within 24 hours, nd who were le to provide written informed consent, were eligile nd recruited sequentilly. Ptients clssified s pllitive y the treting clinicl tem were excluded. Smple Size Clcultion The 10 events per vrile rule ws used to determine the smple size required to study 50 potentil risk fctors. 21,22 This determined tht 1000 ptients, with n verge of 1.5 PIVs ech nd n expected PIV filure of 30% (500 events), were required. Dt Collection At recruitment, seline ptient informtion ws collected y reserch nurse (ReNs) (demogrphics, dmitting dignosis, comoridities, skin type, 23 nd vein condition) nd entered into n electronic dt pltform supported y Reserch Electronic Dt Cpture (REDCp). 24 Bseline dt lso included ctheter vriles (eg, guge, insertion site, ctheterized vein) nd insertion detils (eg, deprtment of insertion, inserting clinicin, numer of insertion ttempts). We included every PIV the prticipnt hd during their dmission until hospitl dischrge or insertion of centrl venous ccess device. PIV sites were reviewed Mondy, Wednesdy, nd Fridy y ReNs for site complictions (eg, redness, pin, swelling, plple cord). Potentil risk fctors for filure were lso recorded (eg, infustes nd dditives, ntiiotic type nd dosge, flushing regimen, numer of times the PIV ws ccessed ech dy for dministrtion of IV medictions or fluids, dressing type nd condition, securement method for the ctheter nd tuing, presence of extension tuing or 3-wy tps, ptient moility sttus, nd delirium). A project mnger trined nd supervised ReNs for protocol complince nd udited study dt qulity. We considered PIV filure to hve occurred if the ctheter hd complictions t removl identified y the ReNs ssessment, from medicl chrts, or y speking to the ptient nd eside nurse. We grouped the filures in 1 of 3 types: (1) occlusion or infiltrtion, defined s lockge, IV fluids moving into surrounding tissue, indurtion, or swelling greter thn 1 cm from the insertion site t or within 24 hours of removl; (2) phleitis, defined s per clinicins definitions or one or more of the following signs nd symptoms: pin or tenderness scored t 2 or more on 1 to 10 incresing severity pin scle, or redness or plple cord (either extending greter thn 1 cm from the insertion site) t or within 24 hours of PIV removl; nd (3) dislodgement (prtil or complete). If multiple complictions were present, ll were recorded. Sttisticl Anlysis Dt were downloded from REDcp to Stt 14.2 (SttCorp., College Sttion, TX) for dt mngement nd nlysis. Missing dt were not imputed. Nominl dt oservtions were collpsed into single oservtion per device. Ptient nd device vriles were descried s frequencies nd proportions, mens nd stndrd devitions, or medins nd interqurtile rnges. Filure incidence rtes were clculted, nd Kpln-Meier survivl curve ws plotted. In generl, Cox proportionl hzrds models were fitted (Efron method) to hndle tied filures (clustering y ptient). Vriles significnt t P < 0.20 on univrile nlyses were sujected to multivrile regression. Generlly, the lrgest ctegory ws set s referent. Correltions etween vriles were checked (Spermn s rnk for inry vriles, R-squred vlue of liner regressions for continuous/ ctegoricl or continuous/continuous vriles). Correltions were considered significnt if r > 0.5 nd the lower ound of the 95% confidence intervl (CI) ws >0.5 (where clculted). Covrite interctions were explored, nd effects t P < 0.05 noted. The 4 steps of multivrile model uilding were (1) seline covrites only with mnul stepwise removl of covrites t P 0.05, (2) tretment covrites only with mnul stepwise removl of covrites t P 0.05, (3) comintion of the derived models from (1) nd (2) nd mnul stepwise removl of covrites t P 0.05, nd (4) mnul stepwise ddition nd removl (t P 0.05) of vriles dropped during the previous steps nd interction testing. Finl models were checked s follows: glol proportionl-hzrds ssumption test, concordnce proility (tht predictions nd outcomes were in greement), nd Nelson-Alen cumultive hzrd function plotted ginst the Cox-Snell residuls. RESULTS Ptient Chrcteristics In totl, 1000 ptients with 1578 PIVs were recruited. The verge ge ws 54 yers nd the mjority were surgicl ptients (673; 67%). Almost hlf of ptients (455; 46%) hd 2 or more comoridities, nd 334 (33%) were oese (ody mss index greter thn 30). Smple chrcteristics re shown y the type of ctheter filure in Tle 1. PIV Chrcteristics All 1578 PIVs were followed until removl, with only 7 PIVs (0.44%) hving missing dt for the 3 outcomes of interest (these were coded s nonfilures for nlysis). Sixty percent of prticipnts hd more thn 1 PIV followed in the study. Doctors nd physicins inserted 1278 (83%) ctheters. A totl of 550 (35%) were plced in the wrd, with 428 (28%) inserted in the emergency deprtment or mulnce. A third of the ctheters (540; 34%) were 18-guge or lrger in dimeter, nd 1000 (64%) were locted in the cuitl foss or hnd. Multiple insertion ttempts were required to plce 315 (23%) PIVs. No PIVs were inserted with ultrsound, s this is rrely used in this hospitl. The flushing policy ws for the dministrtion of 9% sodium chloride every 8 hours if no IV medictions or fluids were ordered. Tle 2 contins further detils of de- E2 An Officil Puliction of the Society of Hospitl Medicine Journl of Hospitl Medicine Pulished Online Octoer 2017

3 Peripherl Intrvenous Ctheter Outcomes Mrsh et l TABLE 1. Prticipnt Chrcteristics t Recruitment y Filure Type Chrcteristic Compliction Clss Totl Occlusion/Infiltrtion Type Phleitis Type Dislodgement Type Group size 1000 (100) 169 (17) 209 (21) 137 (14) Age (y, men nd SD) 54 (19) 57 (17) 52 (19) 59 (19) Sex (mle) 546 (55) 80 (15) 99 (18) 80 (15) BMI (WHO IC) norml or underweight (BMI <25) preoese (25 BMI < 30) oese clss I (30 BMI < 35) oese clss II-III (BMI 35) 301 (30) 358 (36) 182 (18) 152 (15) 43 (14) 66 (18) 25 (14) 35 (23) 65 (22) 71 (20) 37 (20) 33 (22) 50 (17) 45 (13) 23 (13) 18 (12) Skin type: white c 857 (86) 146 (17) 25 (17) 14 (10) Skin integrity d good fir poor 606 (61) 322 (32) 72 (7) 96 (16) 62 (19) 11 (15) 135 (22) 63 (20) 11 (15) 66 (11) 54 (17) 17 (24) Comoridities none or more 259 (26) 286 (29) 181 (18) 136 (14) 138 (14) 30 (12) 49 (17) 38 (21) 52 (20) 66 (23) 37 (20) 28 (20) 24 (9) 34 (12) 28 (15) 25 (18) History of tocco/nicotine use 578 (58) 98 (17) 116 (20) 88 (15) Reson for dmission: gstrointestinl surgery orthopedic surgery vsculr surgery renl surgery other surgery gstrointestinl (medicl) neurology (medicl) other medicl other thn medicl/surgicl 212 (21) 199 (20) 64 (6) 31 (3) 167 (17) 86 (9) 42 (4) 99 (10) 100 (10) 46 (22) 30 (15) 12 (19) 2 (6) 23 (14) 17 (20) 9 (21) 19 (19) 11 (11) 53 (25) 49 (25) 13 (20) 2 (6) 19 (11) 19 (22) 11 (26) 28 (28) 15 (15) 27 (13) 31 (16) 12 (19) 1 (3) 15 (9) 13 (15) 7 (17) 24 (24) 7 (7) Infection (ny type) 107 (11) 34 (32) 43 (40) 17 (16) Wound 502 (50) 83 (17) 99 (20) 68 (14) Drin or IDC 266 (27) 49 (18) 53 (20) 38 (14) No dietry/fluid restrictions 631 (63) 121 (19) 147 (23) 101 (16) Column percentges shown, where pplicle. Row percentges shown, where pplicle. c Skin type 2 s per the Fitzptrick scle. d Good (helthy, well hydrted, elstic), Fir (intct, mildly dehydrted reduced elsticity), Poor (pper, dehydrted-smll mount or no elsticity). NOTE: n (%) shown unless otherwise noted. Occlusion type = occlusion or infiltrtion t removl, indurtion or swelling (>1 cm) t or within ±24 hours of removl; phleitis type = phleitis or pin t removl, pin, tenderness, erythem or plple cord (>1 cm) t or within ±24 hours of removl; dislodgement type = ccidentl removl or dislodgement t removl; proportions clculted using the numer of non-missing vlues in the denomintor; comoridities were s per medicl dignosis documented in ptient s medicl chrt. Arevitions: BMI, ody mss index; IDC, indwelling ctheter; SD, stndrd devition; WHO IC, World Helth Orgniztion Interntionl Clssifiction. vice-relted chrcteristics. Although the hospitl policy ws for ctheter removl y 72 hours, dwell time rnged from <1 to 14 dys, with n verge of 2.4 dys. PIV Complictions Ctheter filure (ny cuse) occurred in 512 (32%) ctheters, which is filure rte of 136 per 1000 ctheter dys (95% CI, ). A totl of 346 ptients out of 1000 (35%) hd t lest 1 filed PIV during the study. Filures were 267 phleitis (17%), 228 occlusion/infiltrtion (14%), nd/or 154 dislodgement (10%; Figure), with some PIVs exhiiting multiple concurrent complictions (Tle 2). Multivrile Anlysis Occlusion/Infiltrtion The multivrile nlysis (Tle 3) showed occlusion or An Officil Puliction of the Society of Hospitl Medicine Journl of Hospitl Medicine Pulished Online Octoer 2017 E3

4 Mrsh et l Peripherl Intrvenous Ctheter Outcomes TABLE 2. Peripherl Intrvenous Ctheter Chrcteristics t Insertion y Filure Type Chrcteristic Compliction Clss Totl Occlusion Type Phleitis Type Dislodgement Type Group size 1578 (100) 228 (14) 267 (17) 154 (10) First device for ptient 1000 (63) 121 (12) 153 (15) 95 (10) Inserted y: doctor nurse prmedic 1278 (83) 170 (11) 92 (6) 177 (14) 35 (21) 14 (15) 218 (17) 29 (17) 11 (12) 118 (9) 14 (8) 16 (17) Inserted t: wrd operting theter emergency deprtment mulnt cre other 550 (35) 481 (30) 340 (22) 88 (6) 119 (8) 106 (19) 42 (9) 51 (15) 14 (16) 15 (13) 113 (21) 57 (12) 67 (20) 10 (11) 20 (17) 61 (11) 31 (6) 33 (10) 16 (18) 13 (11) Size (guge): 16 or lower (13) 346 (23) 733 (49) 217 (15) 15 (8) 40 (12) 104 (14) 50 (23) 24 (12) 58 (17) 127 (17) 43 (20) 19 (10) 27 (8) 61 (8) 32 (15) Loction: hnd cuitl foss lower forerm wrist upper forerm 582 (37) 418 (27) 264 (17) 190 (12) 106 (7) 85 (15) 61 (15) 38 (14) 21 (11) 22 (21) 84 (14) 75 (20) 44 (17) 35 (18) 27 (25) 59 (10) 41 (10) 23 (9) 21 (11) 8 (8) Inserted on dominnt side 738 (47) 114 (15) 151 (20) 76 (10) Difficult insertion c 315 (23) 47 (15) 60 (19) 39 (12) Pin (rnge: 0-10) d 1.9 (2.0) 1.8 (1.9) 2.1 (1.9) 1.7 (1.8) Bruising due to insertion 63 (4) 6 (10) 19 (30) 5 (8) Hir unclipped t insertion 609 (39) 75 (12) 73 (12) 58 (10) Dressing type: Bordered trnsprent dressing simple trnsprent dressing dhesive guze dressing other 742 (47) 592 (38) 134 (9) 97 (6) 128 (17) 71 (12) 13 (10) 12 (12) 143 (19) 87 (15) 17 (13) 19 (20) 86 (12) 48 (8) 8 (6) 11 (11) Column percentges shown, where pplicle. Row percentges shown, where pplicle. c Including multiple insertion ttempts. d Men nd stndrd devition. NOTE: n (%) shown unless otherwise noted; occlusion type = occlusion or infiltrtion t removl, indurtion, or swelling (>1 cm) t or within ±24 hours of removl; phleitis type = phleitis or pin t removl, pin, tenderness, erythem, or plple cord (>1 cm) t or within ±24 hours of removl; dislodgement type = ccidentl removl or dislodgement t removl; proportions clculted y using the numer of nonmissing vlues in the denomintor. infiltrtion ws sttisticlly significntly ssocited with femle ptients (hzrd rtio [HR], 1.48; 95% CI, ), with 22-guge ctheter (HR, 1.43; 95% CI, ), IV flucloxcillin (HR, 1.98; 95% CI, ), nd with frequent PIV ccess (HR, 1.12; 95% CI, ; ie, with ech increse of 1 in the men medictions/fluids dministrtions per dy, reltive PIV filure incresed 112%). Less occlusion nd infiltrtion were sttisticlly significntly ssocited with securement y using dditionl nonsterile tpe (HR, 0.46; 95% CI, ), elsticized tuulr ndges (HR, 0.49; 95% CI, ), or other types of dditionl securement for the PIV (HR, 0.35; 95% CI, ). Phleitis Phleitis ws sttisticlly significntly ssocited with femle ptients (HR, 1.81; 95% CI, ), ruising t the insertion site (HR, 2.16; 95% CI, ), insertion in ptients dominnt side (HR, 1.39; 95% CI, ), IV flucloxicillin (HR, 2.01; 95% CI, ), or with frequent PIV ccess (HR, 1.14; 95% CI, ). Older ge, E4 An Officil Puliction of the Society of Hospitl Medicine Journl of Hospitl Medicine Pulished Online Octoer 2017

5 Peripherl Intrvenous Ctheter Outcomes Mrsh et l TABLE 3. Cox Multivrile Regression y Filure Mode Hzrd Rtios (95% CI) y Compliction Clss Predictors Occlusion/Infiltrtion Type n = 1488 Phleitis Type n = 1565 Dislodgement Type n = 1533 Age (1 y increse) ^ 0.99 ( ) ^ Femle gender 1.48 ( ) 1.81 ( ) ^ Inserted y prmedic (ref. doctor or nurse) ^ ^ 1.78 ( ) 22-guge device 1.43 ( ) ^ ^ Inserted on dominnt side ^ 1.39 ( ) ^ Bruising due to insertion ^ 2.16 ( ) ^ Nonsterile tpe pplied (ever) 0.46 ( ) 0.63 ( ) 0.44 ( ) Elsticized tuulr ndge pplied (ever) 0.49 ( ) ^ ^ Other securement pplied (ever) 0.35 ( ) 0.53 ( ) 0.32 ( ) Numer of ccesses (men per dy) 1.12 ( ) 1.14 ( ) 1.11 ( ) Medictions dministered (t ny time during tril): IV cephzolin IV flucloxcillin ^ 0.63 ( ) ^ 1.98 ( ) 2.01 ( ) ^ P <.001. P <.05. NOTE: Occlusion type = occlusion or infiltrtion t removl, indurtion, or swelling (>1 cm) t or within ±24 hours of removl; phleitis type = phleitis or pin t removl, pin, tenderness, erythem, or plple cord (>1 cm) t or within ±24 hours of removl; dislodgement type = ccidentl removl or dislodgement t removl; ^ not prt of the multivrile model s the results did not rech significnce. Arevitions: CI, confidence intervl; IV, intrvenous; meds, medictions; ref., referent ctegory. (HR, 0.99; 95% CI, ; ie, ech yer older ws ssocited with 1% less phleitis), securement with dditionl nonsterile tpe (HR, 0.63; 95% CI, ) or with ny other dditionl securement (HR, 0.53; 95% CI, ), or the dministrtion of IV cephzolin (HR, 0.63; 95% CI, ) were ssocited with lower phleitis risk. Dislodgement Sttisticlly significnt predictors ssocited with n incresed risk of PIV dislodgement included prmedic insertion (HR, 1.78; 95% CI, ) nd frequent PIV ccess (HR, 1.11; 95% CI, ). A decresed risk ws ssocited with the dditionl securement of the PIV, including nonsterile tpe (HR, 0.44; 95% CI, ) or other forms of dditionl securement (HR, 0.32; 95% CI, ). DISCUSSION One in 3 PIVs filed in this study, with phleitis s the most common cuse of PIV filure. The 17% phleitis rte reflected clinicin-reported phleitis or phleitis oserved y reserch stff using 1-criteri definition ecuse ny sign or symptom cn trigger PIV removl (eg, pin), even if other signs or symptoms re not present. Reported phleitis rtes re lower if definitions require 2 signs or symptoms. 4,6 With over 71 different phleitis ssessment scles in use, nd none well vlidted, the est method for dignosing phleitis remins uncler nd explins the vrition in reported rtes. 25 Occlusion/infiltrtion nd dislodgement were lso highly prevlent forms of PIV filure t 14% nd 10%, respectively. Occlusion nd infiltrtion were comined ecuse clinicl stff use these terms interchngely, nd differentil dignostic tools re not used in prctice. Both result in the sme outcome (therpy interruption nd PIV removl), nd this comintion of outcomes hs een used previously. 23 No PIV-ssocited loodstrem infections occurred, despite the heightened wreness of these infections in the literture Device Dwell Time (Dys) Numer t Risk FIG. Kpln-Meier curve of () occlusion or infiltrtion t removl, () phleitis or pin t removl, nd (c) ccidentl removl or dislodgement. c An Officil Puliction of the Society of Hospitl Medicine Journl of Hospitl Medicine Pulished Online Octoer 2017 E5

6 Mrsh et l Peripherl Intrvenous Ctheter Outcomes Femles hd significntly more occlusion/infiltrtion nd phleitis thn mles, in keeping with previous studies. 7,9,10 This could e ecuse of femles smller vein clier, lthough the effect remined fter djustment for PIV guge. 7,26 The effect of ging on vsculr endothelium nd structurl integrity my explin the oserved decrese in phleitis of 1% with ech older yer of ge. 27 However, gender nd ge effects could e explined y psychosocil fctors (eg, older people my e less likely to dmit pin, or we my question them less symptheticlly), ut, regrdless, women nd younger ptients should e monitored more closely. We found 22-guge ctheters were more likely to fil from occlusion/infiltrtion thn other sizes. This confirms similr findings from Aolfotouh et l. 9 PIV guge selection for this study ws mde t the inserter s discretion nd my e confounded y smller vein size, which ws not mesured. In ddition, risk my e ecuse of smller guge lone or lso more influenced y the shorter length of the studied 22-guge (25 mm) thn the <20-guge ctheters (30 mm). These results question interntionl guidelines, which currently recommend the smllest guge peripherl ctheter possile, 28,29 nd rndomized trils re needed. Although prctice vries etween inserters, some preferentilly cnnulte the nondominnt lim. We re not wre of previous studies on this prctice; however, our results support this pproch. Flucloxcillin ws ssocited with 2-fold increse in occlusion/infiltrtion nd phleitis. Although multiple studies hve reported IV medictions 9,11 nd IV ntiiotics 10,30,31 s risk fctors for PIV filure, none hve identified flucloxcillin s n independent risk fctor. IV flucloxcillin is recommended for reconstitution s 1 g in 15 ml to 20 ml of sterile wter, nd injection over 3 to 4 minutes, lthough this my not e dhered to in prctice. Alterntive dministrtion regimes or improved dherence to current policy my e needed. An exception to the reltionship etween IV ntiiotics nd ctheter filure ws IV cephzolin, ssocited with 40% reltively less phleitis. This my e spurious finding ecuse the dministrtion, ph, nd osmollity of cephzolin re similr to other IV ntiiotics. The more PIVs tht were ccessed per dy, whether for infusions or medictions, the more filure occurred from occlusion/infiltrtion, phleitis, nd dislodgement. This suggests tht peripherl veins re esily dmged nd/or inflmed y the influx of fluids or medictions. Lower injection pressures or the timely trnsfer to orl medictions my limit this prolem. Flushing regimens my lso ssist ecuse prctice vries gretly, nd questions on whether slow continuous flush infusion or intermittent mnul flushing re more vein-protective, nd the optiml flush volume, frequency, nd technique (eg, pulstile) remin. 32,33 Mnul hndling for frequent ccess my loosen dressings nd securement, thus explining the oserved ssocition etween frequent ccess nd ctheter dislodgement. Finlly, the ssocition etween use nd filure my indicte tht mny of these ptients were not suitle for PIV, nd different pproches (eg, ultrsound-guided insertion) or midline my hve een superior option. There is growing emphsis on the need for etter preinsertion ssessment nd selection of the most pproprite device for the ptient nd the IV tretment required. 34 Suoptiml dressings or securements re not unusul in hospitls. 35 Despite our policy of PIV securement with ordered trnsprent dressings, we found 4 dressing types in use. In ddition, we found lmost 50% of PIVs hd n dditionl (secondry) securement, nd this ws ssocited with significntly less PIV filure of ll 3 types. This suggests tht 1 or more of nonsterile tpe, elsticized tuulr ndges, or other securement (eg, ndge or second trnsprent dressing) cn reduce PIV filure, lthough rndomized tril is lcking. 36 Whether the dressing ws filing nd required reinforcement or hospitl stff lcked confidence in the dressing nd plced dditionl securement preventtively is uncler. Both PIV filure nd PIV dressing filure re common, nd further reserch into superior PIV products nd prctices is urgently needed. Prmedic insertions hd higher risk of dislodgement, suggesting tht the incresed emphsis on securement should strt in the prehospitl setting. While multiple or difficult insertion ttempts were not ssocited with PIV filure, insertions were not directly oserved, nd clinicins my hve underreported ttempts. In contrst, insertion-relted ruising ( surrogte for difficult insertion) ws ssocited with more thn doule the incidence of phleitis. The long-term implictions of multiple insertion ttempts on ptient s vsculture re uncler, ut we elieve first time PIV insertion is importnt to ptients nd of interest to clinicins. A recent systemtic review of strtegies ssocited with first ttempt PIV insertion success in n emergency deprtment found little evidence for effective strtegies nd recommended further reserch. 37 The overll PIV filure rte in our study ws 32%, lower thn the 35% to 40% filure oserved in our previous rndomized controlled trils, which hd more stringent inclusion nd exclusion criteri (eg, longer predicted durtion of therpy). 6,38 The implictions for ptients nd costs to the orgniztion of frequent ctheter replcement demonstrte urgent need for further reserch in this re of prctice. 39 A strength of this study is tht ll PIVs, regrdless of the expected length of dwell time or reson for insertion, were eligile for inclusion, providing more generlizle results. The PIV filure rte of 32% is concerning ecuse these filures trigger tretment delys nd replcement insertions, with significnt incresed lor nd equipment costs. The men cost of PIV replcement hs een costed t AUD $69.30 or US $51.92 (s per 2010 $ vlue) per episode of IV tretment. 40 For our hospitl, which uses 200,000 PIVs per yer, the current level of PIV filure suggests lmost AU $5.5 (US $4.1) million in wste nnully t this site lone. The dditionl strengths of this study include the extensive informtion collected prospectively out PIV insertion nd mintennce, including informtion on who inserted the PIV, IV medictions dministered, nd PIV dressings used. Limittions were the popultion of surgicl nd medi- E6 An Officil Puliction of the Society of Hospitl Medicine Journl of Hospitl Medicine Pulished Online Octoer 2017

7 Peripherl Intrvenous Ctheter Outcomes Mrsh et l cl ptients in 1 tertiry hospitl, which my not e generlizle to other settings. CONCLUSION Our study confirms the high rte of ctheter filure in cute cre hospitls, vlidtes existing evidence relted to PIV filure, nd identifies new, potentilly modifile risk fctors to improve PIV insertion nd mngement. Implictions for future reserch were lso identified. Acknowledgments The reserchers cknowledge nd thnk the nurses nd ptients involved in this study. The uthors would lso like to cknowledge Becton Dickinson for prtly funding this study in the form of n unrestricted grnt-in-id pid to Griffith University. Becton Dickinson did not design the study protocol, collect or nlyze dt, nd did not prepre or review the mnuscript. Disclosure: On ehlf of NM nd CMR, Griffith University hs received unrestricted eductionl nd reserch grnts nd consultncy pyment for lectures from 3M nd Becton Dickinson. On ehlf of NM, MC, nd CMR, Griffith University hs received unrestricted investigtor-initited reserch grnts from Centurion Medicl Products nd Entrotech Lifesciences (mnufcturers of PIV dressings) nd Becton Dickinson (mnufcturer of PIVs). On ehlf of MC, Griffith University hs received consultncy pyment to develop eduction mteril from Bxter. On ehlf of CMR, Griffith University hs received unrestricted dontions or investigtor initited reserch grnts unrelted to this reserch from Adhezion, Angiodynmics, Bxter, Crefusion, Cook Medicl, Hospir, Myo, Smiths Medicl, nd Vygon. On ehlf of CMR, Griffith University hs received consultncy pyments for eductionl lectures or professionl opinion from B. Brun, Brd, Crefusion, Myo, ResQDevices, nd Smiths Medicl. On ehlf of EL, Griffith University hs received consultncy pyments for eductionl lecture from 3M. On ehlf of MC, Griffith University hs received consultncy pyment to develop eduction mteril from Bxter. As this ws n oservtionl study, no products were triled in this study. JW nd GM hve no conflicts of interest. References 1. Sri A, Szls J, Holmes KS, Li L, Mussivnd T. Filed ttempts nd improvement strtegies in peripherl intrvenous ctheteriztion. Biomed Mter Eng. 2012;23(1-2): Wester J, Osorne S, Rickrd CM, New K. Cliniclly-indicted replcement versus routine replcement of peripherl venous ctheters. Cochrne Lirry. 2015;8:CD Zingg W, Pittet D. Peripherl venous ctheters: n under-evluted prolem. Int J Antimicro Agents. 2009;34 Suppl 4:S38-S Wester J, Clrke S, Pterson D, et l. Routine cre of peripherl intrvenous ctheters versus cliniclly indicted replcement: rndomised controlled tril. BMJ. 2008;337: Busone-Gzd D, Lefiver CA, Wlters SA. A rndomized controlled tril to compre the complictions of 2 peripherl intrvenous ctheter-stiliztion systems. J Infus Nurs. 2010;33(6): Rickrd CM, Wester J, Wllis MC, et l. Routine versus cliniclly indicted replcement of peripherl intrvenous ctheters: rndomised controlled equivlence tril. Lncet. 2012;380(9847): Dillon MF, Currn J, Mrtos R, et l. Fctors tht ffect longevity of intrvenous cnnuls: prospective study. Q J Med. 2008;101(9): Bolton D. Improving peripherl cnnultion prctice t n NHS Trust. Br J Nurs. 2010;19(21):1346, Aolfotouh MA, Slm M, Bni-Mustf A, White D, Blkhy HH. Prospective study of incidence nd predictors of peripherl intrvenous ctheter-induced complictions. Ther Clin Risk Mng. 2014;10: Wllis MC, McGril M, Wester J, et l. Risk fctors for peripherl intrvenous ctheter filure: multivrite nlysis of dt from rndomized controlled tril. Infect Control Hosp Epidemiol. 2014;35(1): Ctney MR, Hillis S, Wkefield B, et l. Reltionship etween peripherl intrvenous ctheter Dwell time nd the development of phleitis nd infiltrtion. J Infus Nurs. 2001;24(5): Brut F, Pistone T, Guiguet M, et l. Complictions due to peripherl venous ctheteriztion. Prospective study. L Presse Medicle. 2003;32(10): Cicolini G, Bonghi AP, Di Lio L, Di Mscio R. Position of peripherl venous cnnule nd the incidence of thromophleitis: n oservtionl study. J Adv Nurs. 2009;65(6): Sini R, Agnihotri M, Gupt A, Wli I. Epidemiology of infiltrtion nd phleitis. Nursing nd Midwifery Reserch Journl. 2011;7: Plefski SS, Stoddrd GJ. The infusion nurse nd ptient compliction rtes of peripherl-short ctheters: prospective evlution. J Intrven Nurs. 2001;24(2): Bi XH, Zng S, Yu L. A comprison of two intrvenous infusion devices in lung crcinom ptients receiving comined rdiotherpy nd chemotherpy. J Cncer Res Ther. 2013;9(4): Gornsson KE, Johnsson E. Prehospitl peripherl venous ctheters: prospective study of ptient complictions. J Vsc Access. 2012;13(1): Krdeniz G, Kutlu N, Ttlisumk E, Ozkkloglu B. Nurses knowledge regrding ptients with intrvenous ctheters nd phleitis interventions. J Vsc Nurs. 2003;21(2): Fields JM, Den AJ, Todmn RW, et l. The effect of vessel depth, dimeter, nd loction on ultrsound-guided peripherl intrvenous ctheter longevity. Am J Emerg Med. 2012;30(7): McNeill EE, Hines NL, Phriss R. A clinicl tril of new ll-in-one peripherl-short ctheter. J Assoc Vsc Access. 2009;14(1): Vittinghoff E, McCulloch CE. Relxing the rule of ten events per vrile in logistic nd Cox regression. Am J Epidemiol. 2007;165(6): Mllett S, Royston P, Dutton S, Wters R, Altmn DG. Reporting methods in studies developing prognostic models in cncer: review. BMC Med. 2010;8(1): Fitzptrick TB. The vlidity nd prcticlity of sun-rective skin types I through VI. Arch Dermtol. 1988;124(6): Hrris PA, Tylor R, Thielke R, Pyne J, Gonzlez N, Conde JG. Reserch electronic dt cpture (REDCp)-- metdt-driven methodology nd workflow process for providing trnsltionl reserch informtics support. J Biomed Inform. 2009;42(2): Ry-Brruel G, Polit DF, Murfield JE, Rickrd CM. Infusion phleitis ssessment mesures: systemtic review. J Evl Clin Prct. 2014;20(2): Jcoson AF, Winslow EH. Vriles influencing intrvenous ctheter insertion difficulty nd filure: n nlysis of 339 intrvenous ctheter insertions. Hert Lung. 2005;34(5): Schelper R. The ging venous system. Journl of the Assocition for Vsculr Access. 2003;8(3): Gorski L, Hdwy L, Hgle M, McGoldrick M, Orr M, Doellmn D. Infusion therpy stndrds of prctice. J Infus Nurs. 2016;39(suppl 1):S1-S Intrvenous Nursing New Zelnd. Provisionl infusion therpy stndrds of prctice. Accessed Mrch 31, do Rego Furtdo LC. Mintennce of peripherl venous ccess nd its impct on the development of phleitis: survey of 186 ctheters in generl surgery deprtment in Portugl. J Infus Nurs. 2011;34(6): Gorski LA, Hgle ME, Biermn S. Intermittently delivered IV mediction nd ph: reevluting the evidence. J Infus Nurs. 2015;38(1): Keogh S, Flynn J, Mrsh N, Mihl G, Dvies K, Rickrd C. Vried flushing frequency nd volume to prevent peripherl intrvenous ctheter filure: pilot, fctoril rndomised controlled tril in dult medicl-surgicl hospitl ptients. Trils. 2016;17(1): Schreier S, Znchi C, Ronfni L, et l. Norml sline flushes performed once dily mintin peripherl intrvenous ctheter ptency: rndomised controlled tril. Arch Dis Child. 2015;100(7): Chopr V, Flnders SA, Sint S, et l. The Michign Appropriteness Guide for Intrvenous Ctheters (MAGIC): results from multispecilty pnel using the RAND/UCLA ppropriteness method. Ann Intern Med. 2015;163(6 Suppl):S1-S New KA, Wester J, Mrsh NM, Hewer B. Intrvsculr device use, mngement, documenttion nd complictions: point prevlence survey. Aust Helth Rev. 2014;38(3): Mrsh N, Wester J, Mihl G, Rickrd C. Devices nd dressings to secure peripherl venous ctheters to prevent complictions. Cochrne Dtse Syst Rev. 2015(6):CD Prker SI, Benzies KM, Hyden KA, Lng ES. Effectiveness of interventions for dult peripherl intrvenous ctheteriztion: A systemtic review nd met-nlysis of rndomized controlled trils. Int Emerg Nurs. 2016;31: Wester J, Lloyd S, Hopkins T, Osorne S, Yxley M. Developing Reserch se for Intrvenous Peripherl cnnul re-sites (DRIP tril). A rndomised controlled tril of hospitl in-ptients. Int J Nurs Stud. 2007;44(5): Helm RE, Klusner JD, Klemperer JD, Flint LM, Hung E. Accepted ut uncceptle: peripherl IV ctheter filure. J Infus Nurs. 2015;38(3): Tuffh HW, Rickrd CM, Wester J, et l. Cost-effectiveness nlysis of cliniclly indicted versus routine replcement of peripherl intrvenous ctheters. Appl Helth Econ Helth Policy. 2014;12(1): An Officil Puliction of the Society of Hospitl Medicine Journl of Hospitl Medicine Pulished Online Octoer 2017 E7