impact of a clinical decision support system on pharmacy ... · hosp pharm 2013;48(9):1–9 2013...

Hosp Pharm 2013;48(9):1–92013 � Thomas Land Publishers, Inc.www.hospital-pharmacy.comdoi: 10.1310/hpj4809-nnn

Original Article

Impact of a Clinical Decision Support System on PharmacyClinical Interventions, Documentation Efforts, and Costs

Stacy Calloway, MS, PharmD, BCPSp; Hameed A. Akilo, PharmD†; and Kyle Bierman, PharmD, BCPS‡

AbstractHealth care organizations are turning to electronic clinical decision support systems (CDSSs) toincrease quality of patient care and promote a safer environment. A CDSS is a promising approach tothe aggregation and use of patient data to identify patients who would most benefit from interventionsby pharmacy clinicians. However, there are limited published reports describing the impact of CDSSon clinical pharmacy measures. In February 2011, Good Shepherd Medical Center, a 425-bed acutecare community hospital in East Texas, implemented a CDSS (TheraDoc clinical surveillance system).Prior to CDSS implementation, clinicians struggled with obtaining and documenting the dataneeded to support clinical initiatives. The value of having both clinical and staff pharmacists uti-lizing the CDSS has improved communication and knowledge among staff and improved rela-tionships with medical staff, nursing, and case management. The department of pharmacy increasedits clinical interventions from an average of 1,986 per month to 4,065 per month; this representsa 105% increase in the number of interventions. The annual estimated cost savings after CDSS im-plementation is $2,999,508, representing a 96% increase per year and translating into a $1,469,907annual return on investment.

Key Words—antibiotic stewardship, clinical decision support system, costs, documentation, inter-ventions, pharmacy, surveillance

Hosp Pharm—2013;48(9):1–9

The US health care system faces numerous chal-lenges, including the dynamic nature of medicalinformation, clinical practice guidelines, and

mandatory reporting requirements interwoven againsta background of cost containment. New federal re-quirements have been put in place for electronic healthrecords, clinical decision support, and critical patientsafety issues such as the reduction of health care–associated infections, antibiotic-resistant infectious di-seases, and medication errors.

It is difficult to address all these challenges due tothe fragmentation of medical information; data arespread across various databases and systems within anorganization. The use of electronic medical records,computerized order entry, and electronic decision sup-port has been shown to improve the quality of carewithin the hospital environment.1-3 These clinical in-formation systems can be divided into 4 principal

subdomains: notes and records, test results, order entry,and decision support.1 Because these systems are nottypically integrated, hospital-based clinicians are un-able to efficiently retrieve relevant and timely infor-mation that they need to provide quality patient care.

The clinical decision support system (CDSS) system-atically integrates patient data from a variety of hospitalinformation sources to assist clinicians in maximizing thebenefits of clinical surveillance throughout an organiza-tion.4,5 CDSS provides tools for automated, near real-timesurveillance, alerting, analysis, and reporting. Addition-ally, CDSS integrates evidence-based medicine and clinicalguidelines into the delivery of high-quality patient care.4,5

Research has indicated that the use of a CDSS withina hospital system is associated with fewer patient com-plications, lower mortality rates, and lower costs.6

Improving antibiotic use is a key priority for hos-pitals nationwide, and antimicrobial stewardship plays

*Pharmacy Clinical Director; †Critical Care/Oncology; ‡Good Shepherd Medical Center, Longview, Texas. Correspondingauthor: Stacy Calloway, MS, PharmD, BCPS, Pharmacy Clinical Director, Good Shepherd Medical Center, 700 East MarshallAvenue, Longview, TX 75601; phone: 903-658-0329; fax: 903-315-2933; e-mail: [email protected]

Hospital Pharmacy 1

a critical role in fighting antimicrobial resistance byensuring appropriate drug selection, dosing, and du-ration. According to the Centers for Disease Controland Prevention (CDC), almost half of antimicrobialuse in hospitals is unnecessary or inappropriate.7 CDSSintegrates electronic medication administration re-cords with medical administration records, allowinghospitals to monitor antimicrobial utilization. Thistechnology has greatly improved the productivity ofantimicrobial stewardship programs. Combiningavailable technology with expert knowledge in in-fectious diseases is necessary to ensure the continuedefficacy of current antimicrobial agents.8 CDSS canenable pharmacists to compare and leverage patientdata in near real-time from the different systemswithin the hospital, for example, pharmacy and mi-crobiology. CDSS provides tools to aid in the timelyidentification of potential adverse drug events, in-travenous (IV) to oral (PO) conversion opportunities,drug-bug mismatches, and discontinuation or de-es-calation opportunities.

Although CDSS is a highly promising approachfor identifying patients who would most benefit frominterventions by pharmacy clinicians, there are fewreports comparing clinical pharmacy measures pre andpost CDSS implementation. We conducted a compar-ative analysis to identify the impact of this system onpharmacy clinical interventions, clinical documenta-tion, and costs.

METHODSGood Shepherd Medical Center (Good Shepherd)

is a 425-bed acute care community hospital in EastTexas. The pharmacy department uses a decentralizedstaffing model with 7 pharmacists on site, along with 2remote pharmacists during the day shift and 2 duringthe night shift. These pharmacists are decentralized toserve the intensive care unit (ICU), neonatal ICU (NICU),telemetry units, medical/surgery units, central phar-macy, and IV admixture room. Additionally, 2 clinicalpharmacists specializing in oncology/critical care andinternal medicine address complex clinical questionsand perform therapeutic drug monitoring of dangerous,high-risk, and high-cost drugs. These pharmacists fillin staffing roles when needed, attend daily multidis-ciplinary rounds, and perform hospital committee work.

Prior to implementing a CDSS (TheraDoc, Hospira,Lake Forest, IL) in February 2011, Good Shepherddefined institutional business goals relating to the CDSS.These goals included increasing capture and docu-mentation of interventions; determining cost-savingsas a result of pharmacist interventions; improving doc-

umentation of adverse drug reactions; performing DrugUtilization Review and satisfying The Joint Commissionrequirements; identifying and evaluating appropriatelab values pertinent to drug therapy; and performingbenchmark comparisons between institutions.

Set-up and configuration of the new CDSS requireddiscussions and decisions by the team of pharmacyclinicians and management. For example, they decidedwhich clinical equations to load into the CDSS soft-ware. Although the CDSS application gives access tohundreds of alerts, the team chose to begin with just 20alerts, with the staff pharmacists concentrating on thealerts for IV to PO conversion, renal dosing, and van-comycin greater than 72 hours. These alerts coincidedwith the clinical duties that the staff pharmacists werealready performing. The targeted alerts for the clinicalteam were aimed at high-cost drugs and organisms thatwould be especially damaging to patients if left un-treated or treated inappropriately. This introduced theteam to the new workflow and helped them gain con-fidence in and accept the system. Once the staff andinstitution saw positive results with these alerts andhad successfully adopted the new technology, morealerts were added.

The individual institution can stratify alerts into3 tiers based upon the different functionality. Tier 1alerts are considered ‘‘critical alerts’’ and display a redtriangle next to the patient name in a patient search orroster list and in the patient banner. These alerts guidethe clinicians to the patients who should be reviewedfirst in their given patient load. A critical alert may bedefined in different ways by an organization; it couldbe a situation where the pharmacist needs to takeimmediate action to prevent patient harm or it could berelated to key business goals (eg, IV to PO conversion).A Tier 2 alert is a ‘‘daily alert review’’ by function orlocation; this is less critical, but is still very important.A Tier 3 or ‘‘additional alert review’’ identifies con-cerns that may be followed less frequently (eg, weeklyor as time permits). Good Shepherd defined alerts forstaff pharmacists to review each shift and additionaltargeted alerts for the clinical team to review formonthly reporting purposes during the initial imple-mentation of the CDSS (Table 1).

The system was configured to enable alerts to besent to the pharmacists by e-mail and through the hos-pital paging system. A documentation and interven-tion tool that was provided by the CDSS vendor totrack and report interventions (TheraDoc InterventionAssistant) was additionally utilized. The system pro-vides robust intervention review capabilities andallows pharmacy managers to monitor and report

2 Volume 48, October 2013

Impact of CDSS Implementation

interventions by user with customizable time rangeselection. The system documents the exact in-tervention, the total number of interventions, totaltime spent, average time spent, total savings, andaverage savings. This objective data about inter-ventions can be used in employee evaluations. Doc-umentation of clinical interventions is 1 portion ofa 5-tiered annual employee evaluation, with an aver-age of 5 interventions per shift set as standard and anaverage of 8 interventions per shift set as abovestandard. One criterion reads, ‘‘Ensures safe, appro-priate, cost-effective medication therapies for patientsaccording to established policies, procedures, andprotocols.’’ Interventions are reviewed to assesswhether the pharmacist has achieved some of theother domains of this objective, which are as follows:(1) detects and reports suspected adverse drug re-actions and medication errors in a timely manner, (2)assists with pharmacokinetic consult service and drugdosing per organization protocol, (3) provides druginformation and clinical consultation and clarificationto practitioners, (4) sustains the formulary by utilizingtherapeutic substitution protocols, and (5) reviewsand interprets culture and susceptibility data for an-tibiotic appropriateness and recommends changes asneeded.

A standard operating procedure (SOP) is beingcreated at Good Shepherd to ensure that all interven-tions are documented in a consistent and fair mannerand to standardize which activities should be docu-mented as an intervention rather than as regular activi-ties of the job. For example, a renal dosing interventioncan be entered after screening a patient’s profile formedications that need to be adjusted, even if no changeswere made. SOP calls for this to be documented as amedication profile review, which is set at 5 minutes and

$15 cost savings, versus a dose adjustment by phar-macy, which is set at 15 minutes and $75 cost savings. Itis important to review the interventions and maintainthe integrity of the data being collected and reported.SOP discourages a pharmacist from documenting amedication profile review intervention during the courseof order entry, because this is a normal step of properorder entry procedure. Monthly reviews of the systemare conducted at Good Shepherd, and those users withdisproportionately few interventions are encouragedto meet the department goals.

Another program offered by the CDSS vendor isthe Rounds Assistant that is used by clinical phar-macists. A general pharmacy view was designed andis used as a default so that all pharmacists use thesame general pharmacy view. The most importantinformation is identified and included in this view(Figure 1).

Prior to the implementation of the CDSS, phar-macists who attended multidisciplinary rounds wouldperform time-consuming pre-rounding functions, in

Table 1. Alerts defined for initial implementation of TheraDoc for all pharmacists

Daily alert Critical alert

IV to PO conversion Renal function alert

ADR alert Drug level: Antibiotic level

Targeted drug: pipercillin-tazobactam;doripenem, tigecycline, daptomycin,erythropoietins, glycoprotein IIB/IIIAinhibitors, crotalidae polyvalent immuneFAB, linezolid, or vancomycin . 72 hours

Drug level: Phenytoin

TAM: Susceptibility known, inpatient Targeted organism: Pseudomonas, quinolone resistant

TAM: No positive bacterial cultures Targeted organism: Staphylococcus aureus, resistant

TAM: No positive fungal cultures Targeted organism: Enterococcus, vancomycin resistant

Note: ADR 5 adverse drug response; IV to PO 5 intravenous to oral; TAM 5 therapeutic antibiotic monitoring.

Figure 1. Set-up of rounding view information for pharmacists.

Hospital Pharmacy 3


which they would print out patients’ medication pro-files and lists of patients to follow. On a daily basis, 3clinical pharmacists meet on each unit with a multi-disciplinary care team (MDT) consisting of repre-sentatives from social work, case management, dietary,respiratory, nursing, and physical therapy to discussindividual patients in order to optimize their care andexpedite discharge. Prior to the introduction of computer-based services into the MDT, a paper-based record systemwas used to provide case summaries and to documentthe team’s discussion and decisions. These recordscontained free (unstructured) text rather than codedand structured data. The process of recording dis-cussions in an unstructured form (ie, free-text clinicalnotes, scanned documents, pdfs) hindered accuratemeasurement of the MDT performance, as computer-based data analysis and auditing tools cannot be usedon unstructured data.9 There are a number of chal-lenges that the MDT faces in utilizing the CDSS. Theseinclude:

1. Ensuring and documenting adherence with stand-ards (eg, evidence-based guidelines),

2. Identifying patients who are eligible for recruit-ment into clinical trials,

3. Ensuring the consistent collection of crucial data,such as disease staging and outcomes,

4. Establishing robust mechanisms for prospectiveassessment of MDT performance,

5. Ensuring MDT recommendations are followed inpractice,

6. Achieving the right balance of educational andcare delivery objectives of this forum,

7. Establishing reliable interfaces with primary careto ensure continuity of care.

With the implementation of the CDSS, pharmacistsuse tablet computers on rounds. They can access thepatients’ profiles, identify and act on any drug-relatedproblems, and document interventions (Figure 2). Phar-macists can screen patients for opportunities for IV toPO conversions, renal dosing, antibiotic streamlining,and any drug-related problems. While on the units,pharmacists can review the alerts that may have firedfor patients and determine whether to leave a note onthe patients’ chart or call the physician. The use oftablet computers during rounds offers pharmacists themobile platform to document interventions in real-time. This increases the probability of capturing moreinterventions than the practice of taking quick notesand then trying to find time to sit down and enter themafter all other duties have been performed.

Figure 2. Clinical decision support system daily clinicalactivities worksheet.



The use of CDSS provides an instant review of thesusceptibilities of various bacteria to antibiotics in anantibiogram, which can be utilized to identify re-sistance problems. Prior to the availability of on-lineelectronic documentation, the compilation and doc-umentation of antibiogram data was a cumbersomeprocess and required approximately 7 to 14 days tocomplete. This workload was the responsibility of theclinical staff, including the microbiology technician,clinical pharmacist, and clinical pharmacy director.The microbiology technician downloaded the cultureand sensitivity (C&S) report from the prior year intoa worksheet, including information about total iso-lates tested, total patients per isolate, percent of iso-lates susceptible, and resistance. All the isolates werereported for both blood and urine. CDSS implemen-tation allowed simple and easy antibiograms to becreated in less than an hour for the entire institution orany floor within the institution.

Once resistance patterns have been identified, em-piric therapy guidelines specific to the institution canbe developed for various disease states. Informationderived from the antibiogram can also justify the de-velopment of antibiotic restrictions. By running a quickantibiogram, the clinician can ascertain which sus-ceptibilities are highest and choose an empiric treat-ment regimen that may result in improving patientoutcomes by improving time to appropriate therapy.10

One example of this is the development of acommunity-acquired pneumonia (CAP) order set. Byreviewing data from the institution’s antibiogram, theclinical pharmacist can identify preferred antibioticregimens for both non-ICU and ICU patients admittedto the hospital with CAP. A majority of CAP infectionsresult from Streptococcus pneumoniae, Haemophilusinfluenza, Moraxella catarrhalis, and sometimes Es-cherichia coli bacteremia from another source.11 Theclinical pharmacy team performed an antibiogramthat targeted those specific organisms and specific for-mulary drugs that serve as accepted empiric regimensfor non-ICU patients (beta-lactam 1 macrolide) ver-sus an antipseudomonal quinolone. The targeted an-tibiogram led to the team’s decision to make ceftriaxoneplus azithromycin the preferred regimen versus levo-floxacin alone. This recommendation was based onthe increasing evidence in the literature that overuseof quinolones promotes resistance and that E. colisusceptibility to levofloxacin is relatively low at 75%compared to ceftriaxone at 97%.

The CDSS targeted therapeutic antimicrobial mon-itoring alerts are useful for identifying bug-drug mis-matches and positive cultures with no antibiotic orders

on the patients’ profile. Pharmacists are asked to keeptrack of culture results and days of therapy of anti-biotics for the patients on their units to determine whento intervene for de-escalation and discontinuation ofantibiotics. Dose optimization and IV to PO conver-sion play a large role in stewardship efforts. To ensurethat patients are on the proper dose to optimally treatan infection, the dosage of some antibiotics may needto be increased. The renal dosing protocol allows phar-macists to automatically adjust medications, includingcertain antibiotics, to avoid adverse effects in patientswith fluctuations in renal function.

The concept of trigger tools consists of identifyingabnormal lab values or dispensing rescue type medi-cations to ascertain a possible adverse drug event (ADE).For approximately 6 months prior to the implemen-tation of the CDSS system, 2 to 3 pharmacists, a re-presentative from the quality department, and arepresentative from the risk management departmentmet and manually performed a double review of chartsof discharged patients for these triggers using Institutefor Healthcare Improvement Global Trigger Toolmethodology. This review was performed every otherweek, on 50 charts per month, and took approximately1 to 1.5 hours. ADEs were very rarely found duringthese chart review sessions. A better way to target thissearch for ADEs was found through the trigger toolsalert. Investigation of ADEs was targeted and hours ofhigh dollar personnel time was freed up by giving thisfunction to one clinical pharmacist. Lab values in thetrigger tool alert function may include the identifica-tion of abnormalities such as neutropenia, hypogly-cemia, elevated international normalized ratio, elevatedactivated partial thromboplastin time, or elevated serumdrug levels. The use of CDSS to identify these triggersand efficiently compile data in a simplistic and straight-forward fashion makes this program invaluable to thepharmacist when attempting to identify trends of drug-related problems with a certain class of drug or in aparticular unit of the hospital.

The CDSS software has been used to build alertsfor a variety of hospital programs to identify a subsetof the hospital’s patient population. One example isthe diabetes alert, which was built for the diabeteseducator. This alert searches for patients who have labresults for hemoglobin A1C, blood glucose greaterthan 300 mg/dL, or an admit diagnosis of hypoglycemia,hyperglycemia, diabetic ketoacidosis, or hyperglycemichyperosmolar state. This alert identifies newly diag-nosed diabetics and patients who are undiagnosed bytargeting blood glucose readings. The clinician canfollow-up with the appropriate education or a diagnostic

Hospital Pharmacy 5


test recommendation to facilitate improved care of thesepatients.

An alert for drug reimbursement has been built atGood Shepherd to identify all patients who have beenplaced on a particular medication for which the drugrecovery specialist can seek reimbursement, such asamphotericin B, daptomycin, or micafungin. Patientswho may benefit from being placed on standard hos-pital protocols to ensure proper monitoring and treat-ment are also identified. For instance, some patients inthe hospital would develop delirium tremens becausetheir alcohol use was addressed with a onetime orderfor thiamine and folic acid and an ‘‘as needed’’ ben-zodiazepine order. The clinical pharmacy team de-veloped an ‘‘alcohol withdrawal alert’’ that identifiespatients who have been prescribed a benzodiazepine,thiamine, or folic acid; who have had any ethanol/alcohol blood level results; or who have an admit diag-nosis that includes the terms ‘‘alcohol’’ or ‘‘intoxication.’’A recommendation could be made to the physicianthat such patients be placed on the Alcohol With-drawal Protocol. These types of alerts can be tailoredto suit numerous situations, depending on the target atthe time. Alerts can be altered or deactivated when nolonger needed.

RESULTSClinical interventions that were made through the

use of the CDSS had a very high acceptance rate. For

example, in March 2012, there were a total of 4,184interventions made by pharmacists. Of those, 4,172were accepted by physicians, and only 12 were denied.For intervention reporting, interventions can be in-cluded with the general intervention status of (1) any,(2) undetermined, (3) accepted, (4) canceled, (5) ac-cepted modified, or (6) rejected and the follow-up statusof (1) any, (2) pending, or (3) complete. The pharmacydepartment used the intervention reporting structureof completed and accepted, accepted modified, or re-jected. Interventions that were ignored or not actedupon before the patient was discharged were notcounted, as only those interventions that were reportedas completed were captured. Figure 3 breaks down theinterventions by clinical category.

The average number of documented accepted in-terventions by the clinical pharmacy staff increasedfrom 1,986 per month pre CDSS (February 1, 2010,through January 31, 2011) to 4,065 per month postCDSS (February 1, 2012, through January 31, 2013);this represents a 105% increase in the number of in-terventions per year.

Prior to implementing CDSS, cost savings wereroutinely documented as a result of the clinical phar-macy program by using intervention categories andpredetermined cost savings per time and intervention.We used values of time and cost saving from our pre-vious system (Quantifi) to maintain continuity duringthe system change. These numbers are derived from

Figure 3. Clinical intervention breakdown. Anticoag 5 anticoagulation; EPO 5 Epogen; IV to PO 5 intravenous to oral;PK 5 pharmacokinetic; TPN RX 5 total parenteral nutrition prescription; MAR 5 medication administration record.



Table 2. Examples of intervention categories with time (minutes) and estimated cost savings (dollars)

Intervention Description of intervention Time Cost

Pharmacokineticconsult

Initiating therapy per protocol or recommending therapy fordrugs such as aminoglycosides, vancomycin, digoxin, etc.May include additional entries for follow-up when doseadjustments or labs above routine are required.

30 $300

Dose adjustmentby pharmacy

A dose adjustment made by pharmacy based on patient-specificevaluation of clearance, such as renal or hepatic function.The reason for the dose change is to improve efficacy and/orto avoid toxicity. May be a recommendation for dose adjustmentor a protocol-driven adjustment per pharmacy. Requires gatheringand evaluating patient-specific clinical information. There may bemultiple dose adjustments throughout a patient’s stay, so thiscategory may be used more than once per patient.

15 $75

Examples: acyclovir, allopurinol, ampicillin, ampicillin/sulbactam,cefazolin, cefepime, cefoxitin, ceftazadime, digoxin, dofetilide,enoxaparin, fluconazole, levofloxacin, meperidine, meropenem,metformin, metronidazole, nitroprusside, piperacillin/tazobactam,procainamide, propoxyphene, quinidine (same as renal doseevaluation/ change, subtherapeutic dose of frequency,supratherapeutic dose or frequency, and therapeuticrecommendation more appropriate for drug or disease state)

Medicationreconciliation –general

When a pharmacist participates in the reconciliation process(admission, transfer, or discharge) and completes the appropriatereconciliation documents. May involve speaking with the nurseand/or physician.

20 $50

This field allows pharmacist to document a general medicationreconciliation. No medication errors or adverse drug events (ADEs)noted during the process. The focus is on ensuring appropriatecontinuity of care, prevention of omitted medications, and addressingany unmet needs of the patient. Pharmacist may document actualmedication errors and ADEs prevented in the other medicationreconciliation categories in addition to this category.

Medicationreconciliation–major ADE prevented

When a pharmacist prevents a major medication error or ADEdue to participation in the medication reconciliation process(admission, transfer, or discharge). May involve calling homepharmacy or outside physician office and speaking with theinpatient nurse and/or physician.

30 $100

Medicationreconciliation–minor ADE prevented

When a pharmacist prevents a minor medication error or ADEdue to participation in the medication reconciliation process(admission, transfer, or discharge). May involve calling homepharmacy, outside physician office, along with speaking withthe inpatient nurse and/or physician.

15 $100

Therapeuticinterchanges

Therapeutic interchanges: review and implementation 15 $15

Clarify drug order(Same as clarificationof incomplete order)

To be used when patient is interviewed and/or the chart isreviewed to clarify allergy reactions, dosage clarifications,abbreviation clarifications, etc.

10 $50

Examples: You receive an order for Percocet, and an allergyalert to oxycodone appears. You clarify with the patient thatthe "allergy" is only an upset stomach.

Physician orders all home meds to be restarted on the samedoses; you clarify with the patient/family/pharmacy what themedications and doses were at home and write the clarification.

An order for Flovent is missing the strength, so you clarify thestrength with the patient and physician.

(continued)

Hospital Pharmacy 7


an intervention cost calculation model supported bypreviously published pharmacy intervention literature.The model contains variables and assumptions that in-clude interventions that make the medication use moreefficient, promote safety, and promote cost-effectivecare. We adjusted interventions related to direct drugcosts such as IV to PO conversion or discontinuationof inappropriate therapy based on our actual drugcosts. Currently, 77 intervention categories have cal-culated cost savings amounts. Examples are presentedin Table 2.

The total cost savings are calculated on a monthlybasis by adding all the completed and accepted inter-ventions that month and tabulating the associated costsavings. These are tracked and compiled, and 12-monthaverages and projections are made for the pre-CDSSstudy period and the post-CDSS study period.

Cost savings related to pharmacist interventionsincreased from an average of $127,467 per month preCDSS (February 1, 2010, through January 31, 2011)to $249,959 per month post CDSS (February 1, 2012,through January 31, 2013); this represents a 96% in-crease in cost savings per year. There was a $1,469,907annual return on investment, which was calculated bysubtracting the average monthly pre-CDSS interven-tion cost savings from the average monthly post-CDSSintervention cost savings.

DISCUSSIONEighteen months prior to implementing CDSS and

restructuring the clinical program at Good Shepherd,staff struggled with obtaining and documenting thedata they needed to support clinical initiatives. Due tothe ease of use of the software, new hires and phar-macy graduates were able to actively participate in theclinical program within their first few months on staff.The department of pharmacy increased its clinical in-terventions from an average 1,986 per month to 4,065

per month, representing a 105% increase in the numberof interventions per month. The annual estimated costsavings of $2,999,508 post CDSS represents a 96%increase in cost savings per year and translates intoa $1,469,907 annual return on investment.

There are several limitations to this study. We didnot calculate or estimate all of the associated costs withthe implementation of the new software as an expenseagainst our reported savings. We could have attemptedto quantify actual labor cost savings associated withthe workflow optimization strategies described in thearticle. We chose to look at pre- and post-CDSS in-tervention savings only; we collected these data ona monthly basis. We did not collect the data needed tocalculate implementation cost (ie, hours of set up meet-ings, staff training, alert building).

The hospital has made several changes that haveallowed the pharmacy department to increase clinicalinterventions. The hospital started a medical residencyprogram in July 2011, and began daily multidisci-plinary rounds on all units in January 2012. The phar-macy department has been involved in the initiatives ofthe hospital to provide excellent patient care and en-sure evidence-based clinical practice. The increase inthe number of interventions is reflective of the timesaved through workflow optimization with the CDSS.The documentation of clinical interventions by phar-macists demonstrated value and resulted in the ap-proval and hiring of 2 additional full-time clinicalpharmacists; funds have been budgeted for tablet com-puters for pharmacists to use with CDSS on dailymultidisciplinary rounds. One of the new pharmacistpositions specializes in medication safety, which sat-isfies The Joint Commission requirement for hospitalsto designate an individual to carry out the safety re-sponsibilities.12 This individual is responsible for creat-ing reliable team interactions, ensuring reliably designedprocesses, and promoting the value of a just culture.

Table 2. Examples of intervention categories with time (minutes) and estimated cost savings (dollars) (CONT.)

Intervention Description of intervention Time Cost

Code participation To be used for active participation in a code blue. This appliesto neonatal codes as well.

25 $100

Chemotherapyorder review

To be used whenever a chemotherapy order is received andreviewed for protocol, accuracy of doses, organ dysfunction,body surface area calculation, hydration, lifetime doses, etc.Additional categories should be utilized when interventions aremade and antiemetics ordered as a result of this review.

20 $75

Identification ofpatients’ ownmedication

15 $20



CONCLUSIONHealth care organizations and providers face many

challenges associated with the delivery of quality medicalcare. Clinical information technologies, such as elec-tronic medical records, computerized order entry, andelectronic decision support, can improve the quality ofcare within the hospital environment. These systemsintegrate patient data from a variety of hospital in-formation sources and offer tools for automated, nearreal-time surveillance, alerting, analysis, and reporting.At Good Shepherd Medical Center, the CDSS posi-tively impacted pharmacy clinical interventions, clin-ical documentation, and costs.

ACKNOWLEDGMENTSConflict of interest: The authors report that they re-ceived no payment or financial support for this study.Additional contributions: The authors wish to ac-knowledge Anne Gentry, PharmD, for her assistancewith manuscript preparation.

REFERENCES1. Amarasingham R, Diener-West M, Plantinga L, et al.Hospital characteristics associated with highly automated andusable clinical information systems in Texas, United States.BMC Med Inform Decis Mak. 2008;8:39-50.

2. Bates DW. The quality case for information technology inhealthcare. BMC Med Inform Decis Mak. 2002;2:7-10.

3. Bates DW, Gawande AA. Improving safety with informa-tion technology. N Engl J Med. 2003;348(25):2526-2534.

4. Coiera E, ed. Guide to Health Informatics. 2nd ed. London:Arnold; 2003.

5. Kaushal R, Shojania K, Bates D. Effects of computerizedphysician order entry and clinical decision support system onmedication safety: A systemic review. Arch Intern Med. 2003;163:1409-1416.

6. Amarasingham R, Plantinga L, Diener-West M, Gaskin D,Powe N. Clinical information technologies and inpatient out-comes. Arch Intern Med. 2009;169:108-114.

7. Centers for Disease Control and Prevention. Get smart forhealthcare. 2011. http://www.cdc.gov/getsmart/healthcare/.Accessed March 20, 2013.

8. Cannella C. Importance and impact of antimicrobialstewardship. Hosp Pharm. 2010;45(11 Suppl 1):S1-S5.

9. Patkar V, Acosta D, Davidson T, et al. Using comput-erized decision support to improve compliance of cancermultidisciplinary meetings with evidence-based guidance.BMJ Open 2012;2:e000439. doi:10.1136/bmjopen-2011-000439.

10. White Paper. Pogue J. TheraDoc: Incorporating TD intoExisting Antibiotic Stewardship Program. Lake Forest, IL:Hospira; 2011.

11. Howard LS, Sillis M, Pasteur MC, Kamath AV, HarrisonBD. Microbiological profile of community-acquired pneumo-nia in adults over the last 20 years. J Infect. 2005;50(2):107-113.

12. Standard LD.4.40. The leaders ensure that an integratedpatient safety program is implemented throughout the hospi-tal. In: 2006 Comprehensive Accreditation Manual for Hos-pitals: The Official Handbook. Oakbrook Terrace, IL: TheJoint Commission Resources; 2006. g

Hospital Pharmacy 9


P12-0913-0001 Sept., 13

impact of a clinical decision support system on pharmacy ... · hosp pharm 2013;48(9):1–9 2013...

Documents