Table 1 Methodological details and results of studies conducted in intensive care units to reduce inappropriate use of laboratory tests
Article | Réfs. | Years | Study design (Article type) | Setting | Test(s) studied | Intervention(s) type(s) | Time periods | Test(s) reduction or outcome(s) | Costs savings estimate | Safety outcome(s) |
|---|---|---|---|---|---|---|---|---|---|---|
Adhikari et al. | [57] | 2022 | Prospective BAA | N/S | FBC; BMP (n.o.s.) | Education (indications for testing, utilization of results, formal sessions, pamphlets, and flyers) | Pre-I: 2 mo Post-I: 2 mo | Increase in clinically indicated BMP of 19%; no statistical significance for FBC | N/S | N/S |
Aloisio et al. | [81] | 2019 | Retrospective BAA (C.A.) | n = 366 patients | PCT | CPOE (prompts) | Pre-I: 9 mo Post-I: 9 mo | 10% total reduction | EUR 6000 (/9 mo.) | N/S |
Blum et al. | [40] | 2015 | Retrospective BAA | 16-bed; n = 300 patients (159 pre-I; 141 post-I) | ABG | Guidance (literature search) | Pre-I: 3 mo Per-I: 9 mo Post-I: 3 mo | Mean reduction from 5.5 to 3.7 (−38%) ABG tests per patient | N/S | Decrease in MV and LOS |
Bosque et al. | [136] | 2019 | Retrospective BAA (C.A.) | 16-bed; n = 229 PD | N/S | Education (prices information via emails); A&F (number of tests ordered via emails) | Pre-I: 10 mo Per-I: 3 mo Post-I: 9 days | 39% reduction in inappropriate tests for critical patients; no statistical significance for semi-critical patients | N/S | N/S |
Cahill et al. | [41] | 2018 | Retrospective BAA (C.A.) | N/S | N/S | Education (iatrogenic anemia focus culture); Guidance (locally established) | Pre-I: 11 mo Post-I: 11 mo | 23% reduction in laboratory orders; 21% reduction in blood specimens; 23% reduction in POCT specimens | N/S | No increase in LOS nor transfusion need |
Castellanos et al. | [84] | 2018 | Prospective ITS | 25-bed | PCT | CPOE (clinical decision support system implementation) | Pre-I: 4 mo Per-I: 4 consecutives periods of 28 days (ON1-OFF1-ON2-OFF2) Post-I: 28 days | 0.807 TPD on Pre-I (= baseline), 0.662 (−18%) on ON1, 0.733 (−10%/baseline) on OFF1, 0.803 (−0.4%/baseline) on ON2, 0.792 (−2%/baseline) on OFF2, 0.807 (+ 0%/baseline) in Post-I | EUR 15000 (/y) if persistence of scenario “ON1” | N/S |
Cismondi et al. | [98] | 2012 | Database | MIMIC-II database version 2.6; n = 40,426 patients | HCT, HB, PLT, CA, LACT, aPTT, INR/PT, FIB | AI (TS fuzzy modeling; inputs: heart rate, respiratory rate, oxygen saturation, temperature, arterial blood pressure, urine output, intravenous infusions volumes and packed red blood cells, fresh frozen plasma, and platelets transfusions) | N/A | Reduction in 50% of total amount of tests; 11.5% false negatives (= tests that would not be done following algorithm but were in fact appropriate) | N/S | N/A |
Clouzeau et al. | [86] | 2019 | Controlled trial | 2 × 12-bed; n = 5707 patients (3315 interventions; 2392 controls) | CHEM (CREAT, BUN, K, CA, GLU, TBIL, NA, PROT, P, CRP, BNP, PCT, ALT, AST, GGT, TROP); FBC; COAG (n.o.s.), FIB | Education (prices information, formal sessions); A&F (review of tests requested); Gatekeeping (regulation of CA, P, PROT, CREAT, LFT, ABG, TROP testing, proscribe systematic daily electrolytes testing, CRP, systematic check of a first normal electrolyte test) | Pre-I: 1 y Per-I: 1 y. with supervision and 1 y. without supervision Post-I: 7 mo | Per-I (with supervision): −59% total tests/baseline; Per-I (without supervision): −48% total tests/baseline; Post-I: −30%/baseline | EUR 500000 (/y.) | No adverse outcome. No increase in mortality |
Conroy et al. | [80] | 2021 | Prospective BAA | 24-bed | BMP, MG, P, iCA, FBC, COAG (PT, aPTT, INR/PT), TG, CREAT, NH3, pH | Education (posters, new comers, reminders on round checklist); A&F (advice from other clinicians); Gatekeeping (withdrawal of routine admission penal testing); CPOE (warnings minimal retesting interval not respected) | Pre-I: 2 y Per-I: 9 mo | 20% reduction per PD | USD 30000 (/week) | No increase in mortality nor morbidity (RRT, CLABSI, STATs) |
de Bie et al. | [70] | 2016 | Prospective BAA | 48-bed; 9000 PD/y | FBC; COAG (PT/INR, aPTT); CHEM (CREAT, BUN, CA, CL, P, ALB, CRP, AST, ALT, TBIL, ALP, AMY); CARD (CK, CK-MB, TROP) | Education (lectures); Gatekeeping (change in daily morning and post-cardiac surgery routine panels; withdrawal of weekly routine panel constituted of AST, ALT, ALP, AMY, TBIL) CPOE (change in POCT devices presets) | Pre-I: 15 mo Per-I: 15 mo | 24% reduction in total testing | N/S | N/S |
Della-Volpe et al. | [42] | 2014 | Prospective BAA | 71-bed; 22530 ABG/y | ABG | Guidance (locally established following literature review); Education (indications for testing; via email and posters, educational sessions) | Pre-I: 6 weeks Per-I: 6 weeks | Reduction from 35.3 to 26.5 ABGs per 100 PD | USD 87565 (/y.) | N/S |
Dhanani et al. | [74] | 2018 | Prospective BAA | 22-bed; n = 3250 patients (1141 pre-I; 1067 Per-I; 1042 Post-I) | FBC; CHEM (BUN, CREAT, electrolytes, LFT, CA, MG, P); COAG (INR/PT, aPTT, FIB) | Education (emails, posters, weekly staff meetings, scheduled education session, prices information); Guidance (locally established following literature review); CPOE (redesign of request form); A&F (internal audit, daily feedback on tests ordered) | Per-I: 6 mo Per-I: 6 mo Post-I: 6 mo | Per-I: 28% total (FBC 12%, CHEM-20 44%, COAG 70%); Post-I: 26% total | Per-I: reduction of USD 28 (/PD); Post-I: reduction of USD 24 per PD | No increase in mortality nor LOS; no increase in complications (Hb level, MV) |
Dodek et al. | [76] | 2018 | Prospective BAA (C.A.) | 15-bed | MG | Education (new comers to the service, posters); CPOE (prompts); Guidance (locally established following literature review) | Pre-I: 12 mo Post-I: 5 mo | 24% reduction in routine tests with stable non-routine testing | N/S | No increase in mortality nor LOS |
Fresco et al. | [43] | 2016 | Prospective BAA (C.A.) | n = 606 patients (274 pre-I; 342 post-I) | N/S | Guidance (at patient bedside) | Pre-I: 6 mo Post-I: 6 mo | 27% reduction in routine testing | EUR 124000 (/6 mo.) + EUR 53000 (/6 mo.) in transfusion economy | 27% reduction in blood transfusion; no increase in mortality; no difference in nosocomial infections |
Goddard & Austin | [44] | 2011 | Prospective BAA (C.A.) | 6-bed | FBC; CHEM (BUN, electrolytes, LFT, CA, P, ALB, CRP); COAG (n.o.s.) | A&F (audit of personnel via order chart) | Pre-I: 100 days Post-I: 100 days | 33% total reduction (COAG 52%, LFT 54%, BP 54%) | GBP ~ 3000 per bed | N/S |
Gray & Baldwin | [137] | 2014 | Prospective BAA (C.A.) | 26-bed | FBC; CHEM (BUN, electrolytes, LFT, CRP, BP, MG); COAG (n.o.s.) | Education (n.o.s.) | Post-I: 28 days | N/S | EUR ~ 38,000 (/y.) | N/S |
Hall et al. | [138] | 2016 | Prospective BAA (C.A.) | 20-bed; n = 10 patients | N/S | Guidance (clarification of precedent implemented guidelines); Education (posters at patient bedside) | Per-I: 6 random days over a 2 mo. period | Reduction from 46 to 41% in inappropriate tests | N/S | N/S |
Han et al. | [139] | 2014 | Prospective BAA | 600-bed academic hospital (n.o.s.) | iCA, CA, CL, MG, P | Financial incentive; Guidance (locally established) | Pre-I: 1 y Post-I: 1 y | 47% total reduction similarly distributed across all tests | USD 74000 (/y.) and USD 1.7 million in billable charges | No change in quality metrics |
Iosfina et al. | [79] | 2013 | Prospective BAA (C.A.) | 15-bed | FBC, BUN, electrolytes, CREAT | Education (new comers, periodical meetings with clinicians, reminders on checklist); Guidance (locally established); CPOE (prompt with indications for testing) | Pre-I: 1 y Post-I: 3 mo | 24% total reduction | USD 22000 in direct costs | No increase in STAT labs |
Jefferson & King | [68] | 2018 | Prospective BAA | 16-bed; n = 81 patients (41 Pre-I; 40 Post-I) | FBC, CREAT, HCT, LACT, P, PT, TBIL | Education (reminders on checklist, reminders on computers and at bedside); A&F (presence of PI during multidisciplinary rounds to discuss next 24 h tests requests) | Pre-I: 2 weeks Per-I: 2 weeks | NSS | NSS | No statistical differences in mortality and morbidity (reintubation within 48 h, hemorrhage, cardiac arrest, dysrhythmia, hypotension) |
Khan et al. | [75] | 2019 | Prospective BAA (C.A.) | 24-bed | N/S | Education (awareness, visual reminders); CPOE (modification of EMR notes template); A&F (periodical feedback cycles during intervention) | N/S | Reduction in tests from 9.4 to 7.5 (-20%) TPD; increase in discussed clinical cases for testing in morning round from 30 to 95% | N/S | N/S |
Kotecha et al. | [46] | 2017 | Prospective BAA | 12-bed; n = 207 patients (103 Pre-I; 104 Post-I) | COAG (n.o.s.); LFT (n.o.s.), MG, P, LACT, TROP | A&F (audit of testing practice, interview of clinicians); Guidance (locally established); Education (n.o.s.) | Data collected at 2 mo. Post-I; Persistence assessed 1 y. Post-I | Overall reduction of 22% of inappropriate tests; reduction of 39% for LFT testing, 36% COAG, 53% MG, 62% P, 14% CARD, 12% LACT Persistence (/baseline): reduction of 21% for MG, 46% P, 30% LFT, 37% COAG | N/S | No increase in delay of procedures or in cardiac arrhythmias |
Kumwilaisak et al. | [47] | 2008 | Prospective BAA | 25-bed; n = 1117 patients (558 Pre-I; 559 Post-I) | COAG (n.o.s.); CHEM (n.o.s.), GLU, CARD (n.o.s.); ABG | Education (monthly sessions, flyers, emails); Guidance (locally established following literature review); A&F (monthly email on the outcome of the project) | Pre-I: 6 mo Post-I: 6 mo Persistence assessed 1 y. Post-I | 38% reduction on total testing; + 178% increase in tests with clear and appropriate indication; reduction from 21 to 16 (−24%) TPD; decrease in testing of GLU 51%, ABG 48%, CHEM 37%, COG 30%, CARD 23% | N/S | No increase in number of critical values, transfusion need, mortality, LOS, MV, re-admission |
Lee & Maslove | [93] | 2015 | Database | MIMIC-II database version 3.2.09; n = 29,149 patients | HCT, PLT, WBC, GLU, HCO3, K, NA, CL, BUN, CREAT, LACT | Information theory (reviewed in AI-based interventions) | N/A | CREAT-BUN, HCO3-LACT and NA-CL pairs contain a great degree of redundancy of information; PLT, BUN and CREAT are the most redundant tests performed on days 2 and 3 of ICU stay | N/S | N/A |
Leydier et al. | [48] | 2016 | Retrospective BAA (C.A.) | n = 3568 patients | N/S | Guidance (n.o.s.) | Post-I: 5 years | Reduction from 18.1 to 6.4 TPD (−65%) after 5 y | EUR 318000 in 2012 on 5 most expensive tests (n.o.s.) | No increase in transfusion, HB levels, or mortality |
Litton et al. | [87] | 2020 | Prospective BAA | 30-bed; n = 5102 patients (2477 Pre-I; 2625 Post-I) | FBC; CHEM (BUN, electrolytes, MG, CA, LFT (n.o.s.), ABG, CRP, PCT); COAG (INR/PT, aPTT); TROP | Education (relevance of tests); Guidance (n.o.s.); Gatekeeping (withdrawal of routine panel testing and routine admission panel testing); A&F (number of tests during intervention); | Pre-I: 12 mo Post-I: 12 mo | Reduction of routine admission tests from 47.0 to 24.9 (−47%) per ICU admission (= ~ 50,000 in absolute number of tests /y.) | AUD 794000 /y | No increase in mortality nor LOS; reduction of transfusions and MV needed |
Lo et al. | [77] | 2020 | Prospective BAA (L.Ed.) | 15-bed | MG | Education (monthly sessions); CPOE (prompts); Guidance (indications for testing) | Pre-I: 1 y Post-I: 11 mo | Overall reduction from 0.71 to 0.57 (−20%) TPD with reduction from 0.57 to 0.41 (−39%) for routine tests and stable non-routine tests | CAD 4500 /y | No difference in mortality or LOS |
Maguet et al. | [58] | 2015 | Prospective BAA (L.Ed.) | 65-bed; n = 1817 patients (886 Pre-I; 931 Per-I) | FBC; CHEM (n.o.s.); ABG; COAG (PT, aPTT) | Education (daily information, reminders at bedside, prices information) | Pre-I: 4 mo Per-I: 4 mo | Reduction of 7.5% per patient-day | N/A (costs of labs and radiographs mixed) | N/S |
Martinez-B. et al. | [88] | 2017 | Prospective BAA | 98-bed | ABG | Education (classic sessions, posters, stickers on POCT devices, monthly emails); Guidance (locally established following literature review); A&F (audit prior to intervention; monthly feedback emails) | Pre-I: 1 y Post-I: 1 y | 43% | USD 98500 (/y.) | No difference in APACHE, LOS, mortality, MV, re-admission |
Merkeley et al. | [78] | 2016 | Prospective BAA | 15-bed; n = 1440 patients (709 Pre-I; 731 Post-I) | FBC; electrolytes/renal panel | Guidance (locally established following literature review); Education (reminders on morning round checklist, posters, formal sessions); CPOE (prompt on accepted indications) | Pre-I: 1y Post-I: 1 y | Reduction in routine FBC testing from 0.97 to 0.83 TPD (−14%); increase in non-routine FBC testing from 0.37 to 0.40 TPD (+ 8%); Reduction in routine electrolytes testing from 0.96 to 0.83 TPD (−13%); increase in non-routine electrolytes testing from 0.32 to 0.34 (+ 6%) | CAD 11000 (/y.) | No increase in mortality nor complications (number of STAT labs, transfusion, LOS) |
Musca et al. | [90] | 2016 | Controlled trial | 23-bed; n = 253 patients (100 Pre-I; 153 Post-I) | COAG (INR/PT, aPTT, FIB); CHEM (BUN, CREAT, electrolytes) for control | Education (face-to-face, posters, email, electronic material, prices information); Guidance (locally established); A&F (feedback mid-study via email) | Pre-I: 3 mo Post-I: 2 mo | 64% (vs. 15% for control tests) | 60% of total costs (AUD 98000 /y.); AUD ~ 3,8 million to AUS/NZL scale a year | No Increase in complications (bleeding, morbidity) |
Prat et al. | [49] | 2009 | Retrospective BAA | 15-bed; n = 1175 patients (541 Pre-I; 634 Post-I) | COAG (PT, aPTT, FIB); FBC; CHEM (electrolytes, BUN, CREAT, GLU, PROT, CA, P, TBIL, ALP, GGT, AST/ALT, TROP, CK, LACT); ABG | Guidance (locally established following literature review); Education (sessions); A&F (monthly feedback on number of tests) | Pre-I: 1 y Post-I: 1 y | Relative reduction from 38% to 71.5% (overall reduction of ~ 50% of routine tests) | N/A (costs of labs and radiographs mixed) | No difference in mortality nor morbidity (SAPS, MV); decrease in LOS |
Raad et al. | [85] | 2017 | Prospective BAA | 18-bed | N/S | Education (flyers, emails, monthly sessions for resident, during rounds); Guidance (locally established; on urgency of tests); Gatekeeping (withdrawal of routine testing, 24 h max. anticipation for blood analysis prescribing, 1 unique draw at 2 p.m. for tests requested at morning round, withdrawal of ionogram testing on POCT devices); A&F (daily feedback on tests ordered); | Pre-I: 3 mo Post-I: 9 mo | Reduction from 39.4 to 26.9 (−32%) TPD; reduction of patients having daily blood draw from 100 to 12%; reduction of POCT testing from 7.3 to 1.2 (−83%)TPD | N/A (costs of labs and radiographs mixed) | No increase in mortality nor LOS; stable CLABSI |
Rachakonda et al. | [69] | 2017 | Prospective BAA | n ~ 230 patients/mo | CHEM (electrolytes, LFT, CA, P, MG, AMY, LIP, CREAT, BUN); FBC; COAG (n.o.s.); ABG; CARD (CK, CK-MB, TROP); microbiological cultures, microbiological screening swabs, therapeutic drugs level | Education (monthly sessions); A&F (review of tests requested; monthly feedback on the results of the study) | Pre-I: 6 mo Post-I: 6 mo | N/S | Overall reduction in laboratory costs of 12% | No difference in mortality, LOS, nor severity (APACHE III) |
Shen et al. | [50] | 2019 | Controlled trial | 25-bed (46-bed for control) | FBC with differential | A&F (audit of practices and comparison with other hospitals of the same network); Guidance (locally established); Education (n.o.s.) | Pre-I: 2 weeks (2 mo. before intervention) Per-I: 2 weeks Post-I: 2 weeks (2 mo. after intervention) | Reduction in total FBC w/diff Per-I (−20% vs. NSS control) and Post-I (−19% vs. −19% control); reduction in repeated (within 22 h) FBC w/diff Per-I (−31% vs. −21% control) and Post-I (−27% vs. −32% control) | N/S | No negative impact nor delay in the diagnosis of sepsis, no change in management plan |
Simvoulidis et al. | [140] | 2020 | Retrospective BAA (C.A.) | n ~ 1300 patients | N/S | A strategy (n.o.s.) | 1 y | Reduction in > 50% total requests | USD 150000 (/y.) | No negative impact (mortality, LOS, use of invasive resources) |
Sugarman et al. | [71] | 2020 | Prospective BAA (C.A.) | 16-bed; n = 191 PD | FBC, electrolytes, BUN, LFT, CRP, MG, P, COAG (n.o.s.) | Gatekeeping (self-limitation to 25% of requests without clear clinical indication maximum) | Per-I: 4 weeks | Rate of tests requested without clear clinical indication < 25%: CRP (13.1%), FBC (15.4%), BUN & electrolytes (18.1%) Rate of tests requested without clear clinical indication > 25%: LFT (51%), MG (42.2%), P (42.7%), COAG (40.4%) | 25% of costs deemed inappropriate | N/S |
Thakkar et al. | [61] | 2015 | Prospective BAA | 400-bed; n = 1970 | CBC, BMP, CMP, PT/INR, PTT | Educational sessions, flyers, weekly emails with following message: “(1) question the utility of every blood test and order the tests only if the result will affect patient care, (2) think about the sizable impact that costs of blood tests have on health care expenditures, and (3) consider “adding on” tests to blood samples that have already been collected whenever possible” | Pre-I: 2 mo Post-I: 2 mo | Total tests decreased from 13742 pre-I to 13528 post-I (2%) | USD 6.33 per PD | N/S |
Tyrrell et al. | [72] | 2015 | Prospective BAA | 33-bed | BUN, CREAT, electrolytes, BP (CA, ALB, ALP, PROT, calculated globulin), LFT, CA, ALB, MG, P, CRP, FBC, COAG (n.o.s.) | Gatekeeping (implementation of MRI [72 h for BP and LFT; 24 h for CRP] then replaced by SRPT [3 blood draws per week]); Education (trainee medical and nursing staff) | Pre-I: 6 mo Per-I (MRI): 2 periods of 6 mo Per-I (SRPT): 2 periods of 6 mo | 22% total reduction after MRI introduction; 33% total reduction compared to baseline after SRPT introduction | N/S | N/S |
Vezzani et al. | [51] | 2013 | Prospective BAA (L.Ed.) | N/S | CHEM (n.o.s.) | A&F (audit prior to intervention); Education (n.o.s.); Guidance (locally established following literature review) | Pre-I: 1 mo Post-I: 2 mo. separated by 7 mo | 71% decrease in routine testing and 29% in non-routine testing | 37% decrease in costs for routine testing and 63% of costs in non-routine testing | No difference in mortality rate, LOS, nor severity (SAPS II) |
Viau-Lapointe et al. | [91] | 2018 | Prospective BAA (C.A.) | N/S | LFT (n.o.s.); COAG (n.o.s.) | A&F (audit prior to intervention via interview and electronic form); Education (sessions and posters); Guidance (locally established); Gatekeeping (removal of tests deemed unnecessary) | N/S | Reduction for LFT from 0.65 to 0.25 TPD (−61%); NSS for COAG | N/S | N/S |
Walsh et al. | [89] | 2020 | Prospective BAA (C.A.) | 58-bed | ABG | Education (n.o.s.); Guidance (n.o.s.) | Pre-I: 6 mo Post-I: 6 mo | 31% total reduction | AUD 750000 (/y.) | No difference in mortality nor severity (APACHE III) |
Yorkgitis et al. | [60] | 2018 | Controlled trial | 18-bed; n = 307 patients (152 intervention; 155 control) | COAG (PT, aPTT); FBC; CHEM (n.o.s.); ABG | Education (a reminder was added in the checklist round: “What laboratory tests are medically necessary for tomorrow?”; posters) | Pre-I: 3 mo Per-I: 3 mo | NSS | N/A | No difference in mortality, LOS, severity nor morbidity |
Yu et al. | [102] | 2020 | Database | MIMIC-III database; n = 38,773 patients | NA, K, CL, HCO3, CA, MG, P, BUN, CREAT, HB, PLT, WBC | AI (same 2 double-layer Long Short Term Memory Networks modeling as Yu et al.[101] with self-feeding and a corruption strategy; the model had 4 tasks: 1° predict normal vs anormal, 2° predict transition from normal to abnormal or vice-versa, 3° predict numeric value, 4° predict appropriateness; inputs: vitals (n.o.s.), time differences from the last record, race, gender) | N/A | Omissions of 20% of total tests with 98% accuracy (2% false negatives) of abnormality predictions of the omitted tests | N/S | N/A |
Yu et al. | [101] | 2020 | Database | MIMIC-III database; n = 41,113 patients | NA, K, CL, HCO3, CA, MG, P, BUN, CREAT, HB, PLT, WBC | AI (2 double-layer Long Short Term Memory Networks modeling with self-feeding strategy) | N/A | 33% reduction at > 90% accuracy (< 10% false negatives); 15% reduction at > 95% accuracy | N/S | N/A |