Choosing Wisely: Next Steps

Rising healthcare spending has garnered significant public attention, and is considered a threat to other national priorities. Up to one‐third of national health expenditures are wasteful, the largest fraction generated through unnecessary services that could be substituted for less‐costly alternatives or omitted altogether.[1] Physicians play a central role in health spending, as they purchase nearly all tests and therapies on behalf of patients.

One strategy to enhance cost‐conscious physician ordering is to increase transparency of cost data for providers.[2, 3, 4] Although physicians consider price an important factor in ordering decisions, they have difficulty estimating costs accurately or finding price information easily.[5, 6] Improving physicians' knowledge of order costs may prompt them to forego diagnostic tests or therapies of low utility, or shift ordering to lower‐cost alternatives. Real‐time price display during provider order entry is 1 approach for achieving this goal. Modern electronic health records (EHRs) with computerized physician order entry (CPOE) make price display not only practical but also scalable. Integrating price display into clinical workflow, however, can be challenging, and there remains lack of clarity about potential risks and benefits. The dissemination of real‐time CPOE price display, therefore, requires an understanding of its impact on clinical care.

Over the past 3 decades, several studies in the medical literature have evaluated the effect of price display on physician ordering behavior. To date, however, there has been only 1 narrative review of this literature, which did not include several recent studies on the topic or formally address study quality and physician acceptance of price display modules.[7] Therefore, to help inform healthcare leaders, technology innovators, and policy makers, we conducted a systematic review to address 4 key questions: (1) What are the characteristics of interventions that have displayed order prices to physicians in the context of actual practice? (2) To what degree does real‐time display of order prices impact order costs and order volume? (3) Does price display impact patient safety outcomes, and is it acceptable to providers? (4) What is the quality of the current literature on this topic?

METHODS

RESULTS

Database searches yielded a total of 1400 articles, of which 18 were selected on the basis of title and abstract for detailed assessment. Reference searching led us to retrieve 94 further studies of possible interest, of which 23 were selected on the basis of abstract for detailed assessment. Thus, 41 publications underwent full manuscript review, 19 of which met all inclusion criteria (see Supporting Information, Appendix 2, in the online version of this article).[13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31] These studies were published between 1983 and 2014, and were conducted primarily in the United States.

Order Costs and Volume

Thirteen studies reported the numeric impact of price display on aggregate order costs (Table 2). Nine of these demonstrated a statistically significant (P < 0.05) decrease in order costs, with effect sizes ranging from 10.7% to 62.8%.[13, 16, 18, 20, 23, 24, 28, 29, 30] Decreases were found for lab costs, imaging costs, and medication costs, and were observed in both the inpatient and outpatient settings. Three of these 9 studies were randomized. For example, in 1 study randomizing 61 lab tests to price display or no price display, costs for the intervention labs dropped 9.6% compared to the year prior, whereas costs for control labs increased 2.9% (P < 0.001).[16] Two studies randomized by provider group showed that providers seeing order prices accrued 12.7% fewer charges per inpatient admission (P = 0.02) and 12.9% fewer test charges per outpatient visit (P < 0.05).[29, 30] Three studies found no significant association between price display and order costs, with effect sizes ranging from a decrease of 18.8% to an increase of 4.3%.[19, 22, 27] These studies also evaluated lab, imaging, and medication costs, and included 1 randomized trial. One additional large study noted a 12.5% decrease in medication costs after initiation of price display, but did not statistically evaluate this difference.[25]

Table 2.Study Findings

Study	No. of Encounters	Primary Outcome Measure(s)	Impact on Order Costs			Impact on Order Volume
			Control Group Outcome	Intervention Group Outcome	Relative Change	Control Group Outcome	Intervention Group Outcome	Relative Change
NOTE: Abbreviations: ED, emergency department; NA, not applicable; NR, not reported; SICU, surgical intensive care unit.
Fang et al.[14] 2014	378,890 patient‐days	Reference lab orders per 1000 patient‐days	NR	NR	NA	51 orders/1000 patient‐days	38 orders/1000 patient‐days	25.5% orders/1000 patient‐days (P < 0.001)
Nougon et al.[13] 2015	2422 ED visits (excluding washout)	Lab and imaging test costs per ED visit	7.1/visit (lab); 21.8/visit (imaging)	6.4/visit (lab); 14.4/visit (imaging)	10.7% lab costs/ visit (P = 0.02); 33.7% imaging costs/visit (P < 0.001)	NR	NR	NA
Durand et al.[17] 2013	NR	Imaging orders compared to baseline 1 year prior	NR	NR	NA	3.0% total orders	+2.8% total orders	+5.8% total orders (P = 0.10)
Feldman et al.[16] 2013	245,758 patient‐days	Lab orders and fees per patient‐day compared to baseline 1 year prior	+2.9% fees/ patient‐day	9.6% fees/ patient‐day	12.5% fees/patient‐day (P < 0.001)	+5.6% orders/patient‐day	8.6% orders/ patient‐day	14.2% orders/patient‐day (P < 0.001)
Horn et al.[15] 2014	NR	Lab test volume per patient visit, by individual lab test	NR	NR	NA	Aggregate data not reported	Aggregate data not reported	5 of 27 tests had significant reduction in ordering (2.1% to 15.2%/patient visit)
Ellemdin et al.[18] 2011	897 admissions	Lab cost per hospital day	R442.90/day	R284.14/day	35.8% lab costs/patient‐day (P = 0.001)	NR	NR	NA
Schilling[19] 2010	3222 ED visits	Combined lab and imaging test costs per ED visit	108/visit	88/visit	18.8% test costs/visit (P = 0.07)	NR	NR	NA
Guterman et al.[21] 2002	168 urgent care visits	Percent of acid reducer prescriptions for ranitidine (the higher‐cost option)	NR	NR	NA	49% ranitidine	21% ranitidine	57.1% ranitidine (P = 0.007)
Seguin et al.[20] 2002	287 SICU admissions	Tests ordered per admission; test costs per admission	341/admission	266/admission	22.0% test costs/admission (P < 0.05)	13.6 tests/admission	11.1 tests/ admission	18.4% tests/admission (P = 0.12)
Hampers et al.[23] 1999	4881 ED visits (excluding washout)	Adjusted mean test charges per patient visit	$86.79/visit	$63.74/visit	26.6% test charges/visit (P < 0.01)	NR	NR	NA
Ornstein et al.[22] 1999	30,461 outpatient visits	Prescriptions per visit; prescription cost per visit; cost per prescription	$12.49/visit; $21.83/ prescription	$13.03/visit; $22.03/prescription	+4.3% prescription costs/visit (P = 0.12); +0.9% cost/prescription (P = 0.61)	0.66 prescriptions/visit	0.64 prescriptions/ visit	3.0% prescriptions/visit (P value not reported)
Lin et al.[25] 1998	40,747 surgical cases	Annual spending on muscle relaxants medication	$378,234/year (20,389 cases)	$330,923/year (20,358 cases)	12.5%	NR	NR	NA
McNitt et al.[24] 1998	15,130 surgical cases	Anesthesia drug cost per case	$51.02/case	$18.99/case	62.8% drug costs/case (P < 0.05)	NR	NR	NA
Bates et al.[27] 1997	7090 admissions (lab); 17,381 admissions (imaging)	Tests ordered per admission; charges for tests ordered per admission	$771/ admission (lab); $276/admission (imaging)	$739/admission (lab); $275/admission (imaging)	4.2% lab charges/admission (P = 0.97); 0.4% imaging charges/admission (P = 0.10)	26.8 lab tests/admission; 1.76 imaging tests/admission	25.6 lab tests/ admission; 1.76 imaging tests/ admission	4.5% lab tests/admission (P = 0.74); 0% imaging tests/admission (P = 0.13)
Vedsted et al.[26] 1997	NR	Prescribed daily doses per 1000 insured; total drug reimbursement per 1000 insured; reimbursement per daily dose	Reported graphically only	Reported graphically only	No difference	Reported graphically only	Reported graphically only	No difference
Horrow et al.[28] 1994	NR	Anesthetic drugs used per week; anesthetic drug cost per week	$3837/week	$3179/week	17.1% drug costs/week (P = 0.04)	97 drugs/week	94 drugs/week	3.1% drugs/week (P = 0.56)
Tierney et al.[29] 1993	5219 admissions	Total charges per admission	$6964/admission	$6077/admission	12.7% total charges/admission (P = 0.02)	NR	NR	NA
Tierney et al.[30] 1990	15,257 outpatient visits	Test orders per outpatient visit; test charges per outpatient visit	$51.81/visit	$45.13/visit	12.9% test charges/visit (P < 0.05)	1.82 tests/visit	1.56 tests/visit	14.3% tests/visit (P < 0.005)
Everett et al.[31] 1983	NR	Lab tests per admission; charges per admission	NR	NR	NA	NR	NR	No statistically significant changes

Eight studies reported the numeric impact of price display on aggregate order volume. Three of these demonstrated a statistically significant decrease in order volume, with effect sizes ranging from 14.2% to 25.5%.[14, 16, 30] Decreases were found for lab and imaging tests, and were observed in both inpatient and outpatient settings. For example, 1 pre‐post study displaying prices for inpatient send‐out lab tests demonstrated a 25.5% reduction in send‐out labs per 1000 patient‐days (P < 0.001), whereas there was no change for the control group in‐house lab tests, for which prices were not shown.[14] The other 5 studies reported no significant association between price display and order volume, with effect sizes ranging from a decrease of 18.4% to an increase of 5.8%.[17, 20, 22, 27, 28] These studies evaluated lab, imaging, and medication volume. One trial randomizing by individual inpatient showed a nonsignificant decrease of 4.5% in lab orders per admission in the intervention group (P = 0.74), although the authors noted that their study had insufficient power to detect differences less than 10%.[27] Of note, 2 of the 5 studies reporting nonsignificant impacts on order volume (3.1%, P = 0.56; and 18.4%, P = 0.12) did demonstrate significant decreases in order costs (17.1%, P = 0.04; and 22.0%, P < 0.05).[20, 28]

There were an additional 2 studies that reported the impact of price display on order volume for individual orders only. In 1 time‐series study showing lab test prices, there was a statistically significant decrease in order volume for 5 of 27 individual tests studied (using a Bonferroni‐adjusted threshold of significance), with no tests showing a significant increase.[15] In 1 pre‐post study showing prices for H2‐antagonist drugs, there was a statistically significant 57.1% decrease in order volume for the high‐cost medication, with a corresponding 58.7% increase in the low‐cost option.[21] These studies did not report impact on aggregate order costs. Two further studies in this review did not report outcomes numerically, but did state in their articles that significant impacts on order volume were not observed.[26, 31]

Therefore, of the 19 studies included in this review, 17 reported numeric results. Of these 17 studies, 12 showed that price display was associated with statistically significant decreases in either order costs or volume, either in aggregate (10 studies; Figure 1) or for individual orders (2 studies). Of the 7 studies conducted within the past decade, 5 noted significant decreases in order costs or volume. Prices were embedded into an EHR in 5 of these recent studies, and 4 of the 5 observed significant decreases in order costs or volume. Only 2 studies from the past decade1 from Belgium and 1 from the United Statesincorporated prices into an EHR and reported aggregate order costs. Both found statistically significant decreases in order costs with price display.[13, 16]

Figure 1

DISCUSSION

We identified 19 published reports of interventions that displayed real‐time order prices to providers and evaluated the impact on provider ordering. There was substantial heterogeneity in study setting, design, and quality. Although there is insufficient evidence on which to base strong conclusions, these studies collectively suggest that provider price display likely reduces order costs to a modest degree. Data on patient safety were largely lacking, although in the few studies that examined patient outcomes, there was little evidence that patient safety was adversely affected by the intervention. Providers widely viewed display of prices positively.

Our findings align with those of a recent systematic review that concluded that real‐time price information changed provider ordering in the majority of studies.[7] Whereas that review evaluated 17 studies from both clinical settings and simulations, our review focused exclusively on studies conducted in actual ordering environments. Additionally, our literature search yielded 8 studies not previously reviewed. We believe that the alignment of our findings with the prior review, despite the differences in studies included, adds validity to the conclusion that price display likely has a modest impact on reducing order costs. Our review contains several additions important for those considering price display interventions. We provide detailed information on study settings and intervention characteristics. We present a formal assessment of study quality to evaluate the strength of individual study findings and to guide future research in this area. Finally, because both patient safety and provider acceptability may be a concern when prices are shown, we describe all safety outcomes and provider feedback that these studies reported.

The largest effect sizes were noted in 5 studies reporting decreases in order volume or costs greater than 25%.[13, 14, 18, 23, 24] These were all pre‐post intervention studies, so the effect sizes may have been exaggerated by historical confounding. However, the 2 studies with concurrent control groups found no decreases in order volume or cost in the control group.[14, 18] Among the 5 studies that did not find a significant association between price display and provider ordering, 3 were subject to contamination between study groups,[17, 22, 27] 1 was underpowered,[19] and 1 noted a substantial effect size but did not perform a statistical analysis.[25] We also found that order costs were more frequently reduced than order volume, likely because shifts in ordering to less expensive alternatives may cause costs to decrease while volume remains unchanged.[20, 28]

If price display reduces order costs, as the majority of studies in this review indicate, this finding carries broad implications. Policy makers could promote cost‐conscious care by creating incentives for widespread adoption of price display. Hospital and health system leaders could improve transparency and reduce expenses by prioritizing price display. The specific beneficiaries of any reduced spending would depend on payment structures. With shifts toward financial risk‐bearing arrangements like accountable care organizations, healthcare institutions may have a financial interest in adopting price display. Because price display is an administrative intervention that can be developed within EHRs, it is potentially 1 of the most rapidly scalable strategies for reducing healthcare spending. Even modest reductions in spending on laboratory tests, imaging studies, and medications would result in substantial savings on a system‐wide basis.

Implementing price display does not come without challenges. Prices need to be calculated or obtained, loaded into an EHR system, and updated periodically. Technology innovators could enhance EHR software by making these processes easier. Healthcare institutions may find displaying relative prices (eg, $/$$/$$$) logistically simpler in some contexts than showing actual prices (eg, purchase cost), such as when contracts require prices to be confidential. Although we decided to exclude studies displaying relative prices, our search identified no studies that met other inclusion criteria but displayed relative prices, suggesting a lack of evidence regarding the impact of relative price display as an alternative to actual price display.

There are 4 key limitations to our review. First, the heterogeneity of the study designs and reported outcomes precluded pooling of data. The variety of clinical settings and mechanisms through which prices were displayed enhances the generalizability of our findings, but makes it difficult to identify particular contexts (eg, type of price or type of order) in which the intervention may be most effective. Second, although the presence of negative studies on this subject reduces the concern for reporting bias, it remains possible that sites willing to implement and study price displays may be inherently more sensitive to prices, such that published results might be more pronounced than if the intervention were widely implemented across multiple sites. Third, the mixed study quality limits the strength of conclusions that can be drawn. Several studies with both positive and negative findings had issues of bias, contamination, or confounding that make it difficult to be confident of the direction or magnitude of the main findings. Studies evaluating price display are challenging to conduct without these limitations, and that was apparent in our review. Finally, because over half of the studies were conducted over 15 years ago, it may limit their generalizability to modern ordering environments.

We believe there remains a need for high‐quality evidence on this subject within a contemporary context to confirm these findings. The optimal methodology for evaluating this intervention is a cluster randomized trial by facility or provider group, similar to that reported by Tierney et al. in 1990, with a primary outcome of aggregate order costs.[30] Given the substantial investment this would require, a large time series study could also be informative. As most prior price display interventions have been under 6 months in duration, it would be useful to know if the impact on order costs is sustained over a longer time period. The concurrent introduction of any EHR alerts that could impact ordering (eg, duplicate test warnings) should be simultaneously measured and reported. Studies also need to determine the impact of price display alone compared to price comparison displays (displaying prices for the selected order along with reasonable alternatives). Although price comparison was a component of the intervention in some of the studies in this review, it was not evaluated relative to price display alone. Furthermore, it would be helpful to know if the type of price displayed affects its impact. For instance, if providers are most sensitive to the absolute magnitude of prices, then displaying chargemaster prices may impact ordering more than showing hospital costs. If, however, relative prices are all that providers need, then showing lower numbers, such as Medicare prices or hospital costs, may be sufficient. Finally, it would be reassuring to have additional evidence that price display does not adversely impact patient outcomes.

Although some details need elucidation, the studies synthesized in this review provide valuable data in the current climate of increased emphasis on price transparency. Although substantial attention has been devoted by the academic community, technology start‐ups, private insurers, and even state legislatures to improving price transparency to patients, less focus has been given to physicians, for whom healthcare prices are often just as opaque.[4] The findings from this review suggest that provider price display may be an effective, safe, and acceptable approach to empower physicians to control healthcare spending.

Disclosures: Dr. Silvestri, Dr. Bongiovanni, and Ms. Glover have nothing to disclose. Dr. Gross reports grants from Johnson & Johnson, Medtronic Inc., and 21st Century Oncology during the conduct of this study. In addition, he received payment from Fair Health Inc. and ASTRO outside the submitted work.

Rising healthcare spending has garnered significant public attention, and is considered a threat to other national priorities. Up to one‐third of national health expenditures are wasteful, the largest fraction generated through unnecessary services that could be substituted for less‐costly alternatives or omitted altogether.[1] Physicians play a central role in health spending, as they purchase nearly all tests and therapies on behalf of patients.

One strategy to enhance cost‐conscious physician ordering is to increase transparency of cost data for providers.[2, 3, 4] Although physicians consider price an important factor in ordering decisions, they have difficulty estimating costs accurately or finding price information easily.[5, 6] Improving physicians' knowledge of order costs may prompt them to forego diagnostic tests or therapies of low utility, or shift ordering to lower‐cost alternatives. Real‐time price display during provider order entry is 1 approach for achieving this goal. Modern electronic health records (EHRs) with computerized physician order entry (CPOE) make price display not only practical but also scalable. Integrating price display into clinical workflow, however, can be challenging, and there remains lack of clarity about potential risks and benefits. The dissemination of real‐time CPOE price display, therefore, requires an understanding of its impact on clinical care.

Over the past 3 decades, several studies in the medical literature have evaluated the effect of price display on physician ordering behavior. To date, however, there has been only 1 narrative review of this literature, which did not include several recent studies on the topic or formally address study quality and physician acceptance of price display modules.[7] Therefore, to help inform healthcare leaders, technology innovators, and policy makers, we conducted a systematic review to address 4 key questions: (1) What are the characteristics of interventions that have displayed order prices to physicians in the context of actual practice? (2) To what degree does real‐time display of order prices impact order costs and order volume? (3) Does price display impact patient safety outcomes, and is it acceptable to providers? (4) What is the quality of the current literature on this topic?

METHODS

RESULTS

Database searches yielded a total of 1400 articles, of which 18 were selected on the basis of title and abstract for detailed assessment. Reference searching led us to retrieve 94 further studies of possible interest, of which 23 were selected on the basis of abstract for detailed assessment. Thus, 41 publications underwent full manuscript review, 19 of which met all inclusion criteria (see Supporting Information, Appendix 2, in the online version of this article).[13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31] These studies were published between 1983 and 2014, and were conducted primarily in the United States.

Order Costs and Volume

Thirteen studies reported the numeric impact of price display on aggregate order costs (Table 2). Nine of these demonstrated a statistically significant (P < 0.05) decrease in order costs, with effect sizes ranging from 10.7% to 62.8%.[13, 16, 18, 20, 23, 24, 28, 29, 30] Decreases were found for lab costs, imaging costs, and medication costs, and were observed in both the inpatient and outpatient settings. Three of these 9 studies were randomized. For example, in 1 study randomizing 61 lab tests to price display or no price display, costs for the intervention labs dropped 9.6% compared to the year prior, whereas costs for control labs increased 2.9% (P < 0.001).[16] Two studies randomized by provider group showed that providers seeing order prices accrued 12.7% fewer charges per inpatient admission (P = 0.02) and 12.9% fewer test charges per outpatient visit (P < 0.05).[29, 30] Three studies found no significant association between price display and order costs, with effect sizes ranging from a decrease of 18.8% to an increase of 4.3%.[19, 22, 27] These studies also evaluated lab, imaging, and medication costs, and included 1 randomized trial. One additional large study noted a 12.5% decrease in medication costs after initiation of price display, but did not statistically evaluate this difference.[25]

Table 2.Study Findings

Study	No. of Encounters	Primary Outcome Measure(s)	Impact on Order Costs			Impact on Order Volume
			Control Group Outcome	Intervention Group Outcome	Relative Change	Control Group Outcome	Intervention Group Outcome	Relative Change
NOTE: Abbreviations: ED, emergency department; NA, not applicable; NR, not reported; SICU, surgical intensive care unit.
Fang et al.[14] 2014	378,890 patient‐days	Reference lab orders per 1000 patient‐days	NR	NR	NA	51 orders/1000 patient‐days	38 orders/1000 patient‐days	25.5% orders/1000 patient‐days (P < 0.001)
Nougon et al.[13] 2015	2422 ED visits (excluding washout)	Lab and imaging test costs per ED visit	7.1/visit (lab); 21.8/visit (imaging)	6.4/visit (lab); 14.4/visit (imaging)	10.7% lab costs/ visit (P = 0.02); 33.7% imaging costs/visit (P < 0.001)	NR	NR	NA
Durand et al.[17] 2013	NR	Imaging orders compared to baseline 1 year prior	NR	NR	NA	3.0% total orders	+2.8% total orders	+5.8% total orders (P = 0.10)
Feldman et al.[16] 2013	245,758 patient‐days	Lab orders and fees per patient‐day compared to baseline 1 year prior	+2.9% fees/ patient‐day	9.6% fees/ patient‐day	12.5% fees/patient‐day (P < 0.001)	+5.6% orders/patient‐day	8.6% orders/ patient‐day	14.2% orders/patient‐day (P < 0.001)
Horn et al.[15] 2014	NR	Lab test volume per patient visit, by individual lab test	NR	NR	NA	Aggregate data not reported	Aggregate data not reported	5 of 27 tests had significant reduction in ordering (2.1% to 15.2%/patient visit)
Ellemdin et al.[18] 2011	897 admissions	Lab cost per hospital day	R442.90/day	R284.14/day	35.8% lab costs/patient‐day (P = 0.001)	NR	NR	NA
Schilling[19] 2010	3222 ED visits	Combined lab and imaging test costs per ED visit	108/visit	88/visit	18.8% test costs/visit (P = 0.07)	NR	NR	NA
Guterman et al.[21] 2002	168 urgent care visits	Percent of acid reducer prescriptions for ranitidine (the higher‐cost option)	NR	NR	NA	49% ranitidine	21% ranitidine	57.1% ranitidine (P = 0.007)
Seguin et al.[20] 2002	287 SICU admissions	Tests ordered per admission; test costs per admission	341/admission	266/admission	22.0% test costs/admission (P < 0.05)	13.6 tests/admission	11.1 tests/ admission	18.4% tests/admission (P = 0.12)
Hampers et al.[23] 1999	4881 ED visits (excluding washout)	Adjusted mean test charges per patient visit	$86.79/visit	$63.74/visit	26.6% test charges/visit (P < 0.01)	NR	NR	NA
Ornstein et al.[22] 1999	30,461 outpatient visits	Prescriptions per visit; prescription cost per visit; cost per prescription	$12.49/visit; $21.83/ prescription	$13.03/visit; $22.03/prescription	+4.3% prescription costs/visit (P = 0.12); +0.9% cost/prescription (P = 0.61)	0.66 prescriptions/visit	0.64 prescriptions/ visit	3.0% prescriptions/visit (P value not reported)
Lin et al.[25] 1998	40,747 surgical cases	Annual spending on muscle relaxants medication	$378,234/year (20,389 cases)	$330,923/year (20,358 cases)	12.5%	NR	NR	NA
McNitt et al.[24] 1998	15,130 surgical cases	Anesthesia drug cost per case	$51.02/case	$18.99/case	62.8% drug costs/case (P < 0.05)	NR	NR	NA
Bates et al.[27] 1997	7090 admissions (lab); 17,381 admissions (imaging)	Tests ordered per admission; charges for tests ordered per admission	$771/ admission (lab); $276/admission (imaging)	$739/admission (lab); $275/admission (imaging)	4.2% lab charges/admission (P = 0.97); 0.4% imaging charges/admission (P = 0.10)	26.8 lab tests/admission; 1.76 imaging tests/admission	25.6 lab tests/ admission; 1.76 imaging tests/ admission	4.5% lab tests/admission (P = 0.74); 0% imaging tests/admission (P = 0.13)
Vedsted et al.[26] 1997	NR	Prescribed daily doses per 1000 insured; total drug reimbursement per 1000 insured; reimbursement per daily dose	Reported graphically only	Reported graphically only	No difference	Reported graphically only	Reported graphically only	No difference
Horrow et al.[28] 1994	NR	Anesthetic drugs used per week; anesthetic drug cost per week	$3837/week	$3179/week	17.1% drug costs/week (P = 0.04)	97 drugs/week	94 drugs/week	3.1% drugs/week (P = 0.56)
Tierney et al.[29] 1993	5219 admissions	Total charges per admission	$6964/admission	$6077/admission	12.7% total charges/admission (P = 0.02)	NR	NR	NA
Tierney et al.[30] 1990	15,257 outpatient visits	Test orders per outpatient visit; test charges per outpatient visit	$51.81/visit	$45.13/visit	12.9% test charges/visit (P < 0.05)	1.82 tests/visit	1.56 tests/visit	14.3% tests/visit (P < 0.005)
Everett et al.[31] 1983	NR	Lab tests per admission; charges per admission	NR	NR	NA	NR	NR	No statistically significant changes

Eight studies reported the numeric impact of price display on aggregate order volume. Three of these demonstrated a statistically significant decrease in order volume, with effect sizes ranging from 14.2% to 25.5%.[14, 16, 30] Decreases were found for lab and imaging tests, and were observed in both inpatient and outpatient settings. For example, 1 pre‐post study displaying prices for inpatient send‐out lab tests demonstrated a 25.5% reduction in send‐out labs per 1000 patient‐days (P < 0.001), whereas there was no change for the control group in‐house lab tests, for which prices were not shown.[14] The other 5 studies reported no significant association between price display and order volume, with effect sizes ranging from a decrease of 18.4% to an increase of 5.8%.[17, 20, 22, 27, 28] These studies evaluated lab, imaging, and medication volume. One trial randomizing by individual inpatient showed a nonsignificant decrease of 4.5% in lab orders per admission in the intervention group (P = 0.74), although the authors noted that their study had insufficient power to detect differences less than 10%.[27] Of note, 2 of the 5 studies reporting nonsignificant impacts on order volume (3.1%, P = 0.56; and 18.4%, P = 0.12) did demonstrate significant decreases in order costs (17.1%, P = 0.04; and 22.0%, P < 0.05).[20, 28]

There were an additional 2 studies that reported the impact of price display on order volume for individual orders only. In 1 time‐series study showing lab test prices, there was a statistically significant decrease in order volume for 5 of 27 individual tests studied (using a Bonferroni‐adjusted threshold of significance), with no tests showing a significant increase.[15] In 1 pre‐post study showing prices for H2‐antagonist drugs, there was a statistically significant 57.1% decrease in order volume for the high‐cost medication, with a corresponding 58.7% increase in the low‐cost option.[21] These studies did not report impact on aggregate order costs. Two further studies in this review did not report outcomes numerically, but did state in their articles that significant impacts on order volume were not observed.[26, 31]

Therefore, of the 19 studies included in this review, 17 reported numeric results. Of these 17 studies, 12 showed that price display was associated with statistically significant decreases in either order costs or volume, either in aggregate (10 studies; Figure 1) or for individual orders (2 studies). Of the 7 studies conducted within the past decade, 5 noted significant decreases in order costs or volume. Prices were embedded into an EHR in 5 of these recent studies, and 4 of the 5 observed significant decreases in order costs or volume. Only 2 studies from the past decade1 from Belgium and 1 from the United Statesincorporated prices into an EHR and reported aggregate order costs. Both found statistically significant decreases in order costs with price display.[13, 16]

Figure 1

DISCUSSION

We identified 19 published reports of interventions that displayed real‐time order prices to providers and evaluated the impact on provider ordering. There was substantial heterogeneity in study setting, design, and quality. Although there is insufficient evidence on which to base strong conclusions, these studies collectively suggest that provider price display likely reduces order costs to a modest degree. Data on patient safety were largely lacking, although in the few studies that examined patient outcomes, there was little evidence that patient safety was adversely affected by the intervention. Providers widely viewed display of prices positively.

Our findings align with those of a recent systematic review that concluded that real‐time price information changed provider ordering in the majority of studies.[7] Whereas that review evaluated 17 studies from both clinical settings and simulations, our review focused exclusively on studies conducted in actual ordering environments. Additionally, our literature search yielded 8 studies not previously reviewed. We believe that the alignment of our findings with the prior review, despite the differences in studies included, adds validity to the conclusion that price display likely has a modest impact on reducing order costs. Our review contains several additions important for those considering price display interventions. We provide detailed information on study settings and intervention characteristics. We present a formal assessment of study quality to evaluate the strength of individual study findings and to guide future research in this area. Finally, because both patient safety and provider acceptability may be a concern when prices are shown, we describe all safety outcomes and provider feedback that these studies reported.

The largest effect sizes were noted in 5 studies reporting decreases in order volume or costs greater than 25%.[13, 14, 18, 23, 24] These were all pre‐post intervention studies, so the effect sizes may have been exaggerated by historical confounding. However, the 2 studies with concurrent control groups found no decreases in order volume or cost in the control group.[14, 18] Among the 5 studies that did not find a significant association between price display and provider ordering, 3 were subject to contamination between study groups,[17, 22, 27] 1 was underpowered,[19] and 1 noted a substantial effect size but did not perform a statistical analysis.[25] We also found that order costs were more frequently reduced than order volume, likely because shifts in ordering to less expensive alternatives may cause costs to decrease while volume remains unchanged.[20, 28]

If price display reduces order costs, as the majority of studies in this review indicate, this finding carries broad implications. Policy makers could promote cost‐conscious care by creating incentives for widespread adoption of price display. Hospital and health system leaders could improve transparency and reduce expenses by prioritizing price display. The specific beneficiaries of any reduced spending would depend on payment structures. With shifts toward financial risk‐bearing arrangements like accountable care organizations, healthcare institutions may have a financial interest in adopting price display. Because price display is an administrative intervention that can be developed within EHRs, it is potentially 1 of the most rapidly scalable strategies for reducing healthcare spending. Even modest reductions in spending on laboratory tests, imaging studies, and medications would result in substantial savings on a system‐wide basis.

Implementing price display does not come without challenges. Prices need to be calculated or obtained, loaded into an EHR system, and updated periodically. Technology innovators could enhance EHR software by making these processes easier. Healthcare institutions may find displaying relative prices (eg, $/$$/$$$) logistically simpler in some contexts than showing actual prices (eg, purchase cost), such as when contracts require prices to be confidential. Although we decided to exclude studies displaying relative prices, our search identified no studies that met other inclusion criteria but displayed relative prices, suggesting a lack of evidence regarding the impact of relative price display as an alternative to actual price display.

There are 4 key limitations to our review. First, the heterogeneity of the study designs and reported outcomes precluded pooling of data. The variety of clinical settings and mechanisms through which prices were displayed enhances the generalizability of our findings, but makes it difficult to identify particular contexts (eg, type of price or type of order) in which the intervention may be most effective. Second, although the presence of negative studies on this subject reduces the concern for reporting bias, it remains possible that sites willing to implement and study price displays may be inherently more sensitive to prices, such that published results might be more pronounced than if the intervention were widely implemented across multiple sites. Third, the mixed study quality limits the strength of conclusions that can be drawn. Several studies with both positive and negative findings had issues of bias, contamination, or confounding that make it difficult to be confident of the direction or magnitude of the main findings. Studies evaluating price display are challenging to conduct without these limitations, and that was apparent in our review. Finally, because over half of the studies were conducted over 15 years ago, it may limit their generalizability to modern ordering environments.

We believe there remains a need for high‐quality evidence on this subject within a contemporary context to confirm these findings. The optimal methodology for evaluating this intervention is a cluster randomized trial by facility or provider group, similar to that reported by Tierney et al. in 1990, with a primary outcome of aggregate order costs.[30] Given the substantial investment this would require, a large time series study could also be informative. As most prior price display interventions have been under 6 months in duration, it would be useful to know if the impact on order costs is sustained over a longer time period. The concurrent introduction of any EHR alerts that could impact ordering (eg, duplicate test warnings) should be simultaneously measured and reported. Studies also need to determine the impact of price display alone compared to price comparison displays (displaying prices for the selected order along with reasonable alternatives). Although price comparison was a component of the intervention in some of the studies in this review, it was not evaluated relative to price display alone. Furthermore, it would be helpful to know if the type of price displayed affects its impact. For instance, if providers are most sensitive to the absolute magnitude of prices, then displaying chargemaster prices may impact ordering more than showing hospital costs. If, however, relative prices are all that providers need, then showing lower numbers, such as Medicare prices or hospital costs, may be sufficient. Finally, it would be reassuring to have additional evidence that price display does not adversely impact patient outcomes.

Although some details need elucidation, the studies synthesized in this review provide valuable data in the current climate of increased emphasis on price transparency. Although substantial attention has been devoted by the academic community, technology start‐ups, private insurers, and even state legislatures to improving price transparency to patients, less focus has been given to physicians, for whom healthcare prices are often just as opaque.[4] The findings from this review suggest that provider price display may be an effective, safe, and acceptable approach to empower physicians to control healthcare spending.

Disclosures: Dr. Silvestri, Dr. Bongiovanni, and Ms. Glover have nothing to disclose. Dr. Gross reports grants from Johnson & Johnson, Medtronic Inc., and 21st Century Oncology during the conduct of this study. In addition, he received payment from Fair Health Inc. and ASTRO outside the submitted work.

Rising healthcare spending has garnered significant public attention, and is considered a threat to other national priorities. Up to one‐third of national health expenditures are wasteful, the largest fraction generated through unnecessary services that could be substituted for less‐costly alternatives or omitted altogether.[1] Physicians play a central role in health spending, as they purchase nearly all tests and therapies on behalf of patients.

One strategy to enhance cost‐conscious physician ordering is to increase transparency of cost data for providers.[2, 3, 4] Although physicians consider price an important factor in ordering decisions, they have difficulty estimating costs accurately or finding price information easily.[5, 6] Improving physicians' knowledge of order costs may prompt them to forego diagnostic tests or therapies of low utility, or shift ordering to lower‐cost alternatives. Real‐time price display during provider order entry is 1 approach for achieving this goal. Modern electronic health records (EHRs) with computerized physician order entry (CPOE) make price display not only practical but also scalable. Integrating price display into clinical workflow, however, can be challenging, and there remains lack of clarity about potential risks and benefits. The dissemination of real‐time CPOE price display, therefore, requires an understanding of its impact on clinical care.

Over the past 3 decades, several studies in the medical literature have evaluated the effect of price display on physician ordering behavior. To date, however, there has been only 1 narrative review of this literature, which did not include several recent studies on the topic or formally address study quality and physician acceptance of price display modules.[7] Therefore, to help inform healthcare leaders, technology innovators, and policy makers, we conducted a systematic review to address 4 key questions: (1) What are the characteristics of interventions that have displayed order prices to physicians in the context of actual practice? (2) To what degree does real‐time display of order prices impact order costs and order volume? (3) Does price display impact patient safety outcomes, and is it acceptable to providers? (4) What is the quality of the current literature on this topic?

METHODS

RESULTS

Database searches yielded a total of 1400 articles, of which 18 were selected on the basis of title and abstract for detailed assessment. Reference searching led us to retrieve 94 further studies of possible interest, of which 23 were selected on the basis of abstract for detailed assessment. Thus, 41 publications underwent full manuscript review, 19 of which met all inclusion criteria (see Supporting Information, Appendix 2, in the online version of this article).[13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31] These studies were published between 1983 and 2014, and were conducted primarily in the United States.

Order Costs and Volume

Thirteen studies reported the numeric impact of price display on aggregate order costs (Table 2). Nine of these demonstrated a statistically significant (P < 0.05) decrease in order costs, with effect sizes ranging from 10.7% to 62.8%.[13, 16, 18, 20, 23, 24, 28, 29, 30] Decreases were found for lab costs, imaging costs, and medication costs, and were observed in both the inpatient and outpatient settings. Three of these 9 studies were randomized. For example, in 1 study randomizing 61 lab tests to price display or no price display, costs for the intervention labs dropped 9.6% compared to the year prior, whereas costs for control labs increased 2.9% (P < 0.001).[16] Two studies randomized by provider group showed that providers seeing order prices accrued 12.7% fewer charges per inpatient admission (P = 0.02) and 12.9% fewer test charges per outpatient visit (P < 0.05).[29, 30] Three studies found no significant association between price display and order costs, with effect sizes ranging from a decrease of 18.8% to an increase of 4.3%.[19, 22, 27] These studies also evaluated lab, imaging, and medication costs, and included 1 randomized trial. One additional large study noted a 12.5% decrease in medication costs after initiation of price display, but did not statistically evaluate this difference.[25]

Table 2.Study Findings

Study	No. of Encounters	Primary Outcome Measure(s)	Impact on Order Costs			Impact on Order Volume
			Control Group Outcome	Intervention Group Outcome	Relative Change	Control Group Outcome	Intervention Group Outcome	Relative Change
NOTE: Abbreviations: ED, emergency department; NA, not applicable; NR, not reported; SICU, surgical intensive care unit.
Fang et al.[14] 2014	378,890 patient‐days	Reference lab orders per 1000 patient‐days	NR	NR	NA	51 orders/1000 patient‐days	38 orders/1000 patient‐days	25.5% orders/1000 patient‐days (P < 0.001)
Nougon et al.[13] 2015	2422 ED visits (excluding washout)	Lab and imaging test costs per ED visit	7.1/visit (lab); 21.8/visit (imaging)	6.4/visit (lab); 14.4/visit (imaging)	10.7% lab costs/ visit (P = 0.02); 33.7% imaging costs/visit (P < 0.001)	NR	NR	NA
Durand et al.[17] 2013	NR	Imaging orders compared to baseline 1 year prior	NR	NR	NA	3.0% total orders	+2.8% total orders	+5.8% total orders (P = 0.10)
Feldman et al.[16] 2013	245,758 patient‐days	Lab orders and fees per patient‐day compared to baseline 1 year prior	+2.9% fees/ patient‐day	9.6% fees/ patient‐day	12.5% fees/patient‐day (P < 0.001)	+5.6% orders/patient‐day	8.6% orders/ patient‐day	14.2% orders/patient‐day (P < 0.001)
Horn et al.[15] 2014	NR	Lab test volume per patient visit, by individual lab test	NR	NR	NA	Aggregate data not reported	Aggregate data not reported	5 of 27 tests had significant reduction in ordering (2.1% to 15.2%/patient visit)
Ellemdin et al.[18] 2011	897 admissions	Lab cost per hospital day	R442.90/day	R284.14/day	35.8% lab costs/patient‐day (P = 0.001)	NR	NR	NA
Schilling[19] 2010	3222 ED visits	Combined lab and imaging test costs per ED visit	108/visit	88/visit	18.8% test costs/visit (P = 0.07)	NR	NR	NA
Guterman et al.[21] 2002	168 urgent care visits	Percent of acid reducer prescriptions for ranitidine (the higher‐cost option)	NR	NR	NA	49% ranitidine	21% ranitidine	57.1% ranitidine (P = 0.007)
Seguin et al.[20] 2002	287 SICU admissions	Tests ordered per admission; test costs per admission	341/admission	266/admission	22.0% test costs/admission (P < 0.05)	13.6 tests/admission	11.1 tests/ admission	18.4% tests/admission (P = 0.12)
Hampers et al.[23] 1999	4881 ED visits (excluding washout)	Adjusted mean test charges per patient visit	$86.79/visit	$63.74/visit	26.6% test charges/visit (P < 0.01)	NR	NR	NA
Ornstein et al.[22] 1999	30,461 outpatient visits	Prescriptions per visit; prescription cost per visit; cost per prescription	$12.49/visit; $21.83/ prescription	$13.03/visit; $22.03/prescription	+4.3% prescription costs/visit (P = 0.12); +0.9% cost/prescription (P = 0.61)	0.66 prescriptions/visit	0.64 prescriptions/ visit	3.0% prescriptions/visit (P value not reported)
Lin et al.[25] 1998	40,747 surgical cases	Annual spending on muscle relaxants medication	$378,234/year (20,389 cases)	$330,923/year (20,358 cases)	12.5%	NR	NR	NA
McNitt et al.[24] 1998	15,130 surgical cases	Anesthesia drug cost per case	$51.02/case	$18.99/case	62.8% drug costs/case (P < 0.05)	NR	NR	NA
Bates et al.[27] 1997	7090 admissions (lab); 17,381 admissions (imaging)	Tests ordered per admission; charges for tests ordered per admission	$771/ admission (lab); $276/admission (imaging)	$739/admission (lab); $275/admission (imaging)	4.2% lab charges/admission (P = 0.97); 0.4% imaging charges/admission (P = 0.10)	26.8 lab tests/admission; 1.76 imaging tests/admission	25.6 lab tests/ admission; 1.76 imaging tests/ admission	4.5% lab tests/admission (P = 0.74); 0% imaging tests/admission (P = 0.13)
Vedsted et al.[26] 1997	NR	Prescribed daily doses per 1000 insured; total drug reimbursement per 1000 insured; reimbursement per daily dose	Reported graphically only	Reported graphically only	No difference	Reported graphically only	Reported graphically only	No difference
Horrow et al.[28] 1994	NR	Anesthetic drugs used per week; anesthetic drug cost per week	$3837/week	$3179/week	17.1% drug costs/week (P = 0.04)	97 drugs/week	94 drugs/week	3.1% drugs/week (P = 0.56)
Tierney et al.[29] 1993	5219 admissions	Total charges per admission	$6964/admission	$6077/admission	12.7% total charges/admission (P = 0.02)	NR	NR	NA
Tierney et al.[30] 1990	15,257 outpatient visits	Test orders per outpatient visit; test charges per outpatient visit	$51.81/visit	$45.13/visit	12.9% test charges/visit (P < 0.05)	1.82 tests/visit	1.56 tests/visit	14.3% tests/visit (P < 0.005)
Everett et al.[31] 1983	NR	Lab tests per admission; charges per admission	NR	NR	NA	NR	NR	No statistically significant changes

Eight studies reported the numeric impact of price display on aggregate order volume. Three of these demonstrated a statistically significant decrease in order volume, with effect sizes ranging from 14.2% to 25.5%.[14, 16, 30] Decreases were found for lab and imaging tests, and were observed in both inpatient and outpatient settings. For example, 1 pre‐post study displaying prices for inpatient send‐out lab tests demonstrated a 25.5% reduction in send‐out labs per 1000 patient‐days (P < 0.001), whereas there was no change for the control group in‐house lab tests, for which prices were not shown.[14] The other 5 studies reported no significant association between price display and order volume, with effect sizes ranging from a decrease of 18.4% to an increase of 5.8%.[17, 20, 22, 27, 28] These studies evaluated lab, imaging, and medication volume. One trial randomizing by individual inpatient showed a nonsignificant decrease of 4.5% in lab orders per admission in the intervention group (P = 0.74), although the authors noted that their study had insufficient power to detect differences less than 10%.[27] Of note, 2 of the 5 studies reporting nonsignificant impacts on order volume (3.1%, P = 0.56; and 18.4%, P = 0.12) did demonstrate significant decreases in order costs (17.1%, P = 0.04; and 22.0%, P < 0.05).[20, 28]

There were an additional 2 studies that reported the impact of price display on order volume for individual orders only. In 1 time‐series study showing lab test prices, there was a statistically significant decrease in order volume for 5 of 27 individual tests studied (using a Bonferroni‐adjusted threshold of significance), with no tests showing a significant increase.[15] In 1 pre‐post study showing prices for H2‐antagonist drugs, there was a statistically significant 57.1% decrease in order volume for the high‐cost medication, with a corresponding 58.7% increase in the low‐cost option.[21] These studies did not report impact on aggregate order costs. Two further studies in this review did not report outcomes numerically, but did state in their articles that significant impacts on order volume were not observed.[26, 31]

Therefore, of the 19 studies included in this review, 17 reported numeric results. Of these 17 studies, 12 showed that price display was associated with statistically significant decreases in either order costs or volume, either in aggregate (10 studies; Figure 1) or for individual orders (2 studies). Of the 7 studies conducted within the past decade, 5 noted significant decreases in order costs or volume. Prices were embedded into an EHR in 5 of these recent studies, and 4 of the 5 observed significant decreases in order costs or volume. Only 2 studies from the past decade1 from Belgium and 1 from the United Statesincorporated prices into an EHR and reported aggregate order costs. Both found statistically significant decreases in order costs with price display.[13, 16]

Figure 1

DISCUSSION

We identified 19 published reports of interventions that displayed real‐time order prices to providers and evaluated the impact on provider ordering. There was substantial heterogeneity in study setting, design, and quality. Although there is insufficient evidence on which to base strong conclusions, these studies collectively suggest that provider price display likely reduces order costs to a modest degree. Data on patient safety were largely lacking, although in the few studies that examined patient outcomes, there was little evidence that patient safety was adversely affected by the intervention. Providers widely viewed display of prices positively.

Our findings align with those of a recent systematic review that concluded that real‐time price information changed provider ordering in the majority of studies.[7] Whereas that review evaluated 17 studies from both clinical settings and simulations, our review focused exclusively on studies conducted in actual ordering environments. Additionally, our literature search yielded 8 studies not previously reviewed. We believe that the alignment of our findings with the prior review, despite the differences in studies included, adds validity to the conclusion that price display likely has a modest impact on reducing order costs. Our review contains several additions important for those considering price display interventions. We provide detailed information on study settings and intervention characteristics. We present a formal assessment of study quality to evaluate the strength of individual study findings and to guide future research in this area. Finally, because both patient safety and provider acceptability may be a concern when prices are shown, we describe all safety outcomes and provider feedback that these studies reported.

The largest effect sizes were noted in 5 studies reporting decreases in order volume or costs greater than 25%.[13, 14, 18, 23, 24] These were all pre‐post intervention studies, so the effect sizes may have been exaggerated by historical confounding. However, the 2 studies with concurrent control groups found no decreases in order volume or cost in the control group.[14, 18] Among the 5 studies that did not find a significant association between price display and provider ordering, 3 were subject to contamination between study groups,[17, 22, 27] 1 was underpowered,[19] and 1 noted a substantial effect size but did not perform a statistical analysis.[25] We also found that order costs were more frequently reduced than order volume, likely because shifts in ordering to less expensive alternatives may cause costs to decrease while volume remains unchanged.[20, 28]

If price display reduces order costs, as the majority of studies in this review indicate, this finding carries broad implications. Policy makers could promote cost‐conscious care by creating incentives for widespread adoption of price display. Hospital and health system leaders could improve transparency and reduce expenses by prioritizing price display. The specific beneficiaries of any reduced spending would depend on payment structures. With shifts toward financial risk‐bearing arrangements like accountable care organizations, healthcare institutions may have a financial interest in adopting price display. Because price display is an administrative intervention that can be developed within EHRs, it is potentially 1 of the most rapidly scalable strategies for reducing healthcare spending. Even modest reductions in spending on laboratory tests, imaging studies, and medications would result in substantial savings on a system‐wide basis.

Implementing price display does not come without challenges. Prices need to be calculated or obtained, loaded into an EHR system, and updated periodically. Technology innovators could enhance EHR software by making these processes easier. Healthcare institutions may find displaying relative prices (eg, $/$$/$$$) logistically simpler in some contexts than showing actual prices (eg, purchase cost), such as when contracts require prices to be confidential. Although we decided to exclude studies displaying relative prices, our search identified no studies that met other inclusion criteria but displayed relative prices, suggesting a lack of evidence regarding the impact of relative price display as an alternative to actual price display.

There are 4 key limitations to our review. First, the heterogeneity of the study designs and reported outcomes precluded pooling of data. The variety of clinical settings and mechanisms through which prices were displayed enhances the generalizability of our findings, but makes it difficult to identify particular contexts (eg, type of price or type of order) in which the intervention may be most effective. Second, although the presence of negative studies on this subject reduces the concern for reporting bias, it remains possible that sites willing to implement and study price displays may be inherently more sensitive to prices, such that published results might be more pronounced than if the intervention were widely implemented across multiple sites. Third, the mixed study quality limits the strength of conclusions that can be drawn. Several studies with both positive and negative findings had issues of bias, contamination, or confounding that make it difficult to be confident of the direction or magnitude of the main findings. Studies evaluating price display are challenging to conduct without these limitations, and that was apparent in our review. Finally, because over half of the studies were conducted over 15 years ago, it may limit their generalizability to modern ordering environments.

We believe there remains a need for high‐quality evidence on this subject within a contemporary context to confirm these findings. The optimal methodology for evaluating this intervention is a cluster randomized trial by facility or provider group, similar to that reported by Tierney et al. in 1990, with a primary outcome of aggregate order costs.[30] Given the substantial investment this would require, a large time series study could also be informative. As most prior price display interventions have been under 6 months in duration, it would be useful to know if the impact on order costs is sustained over a longer time period. The concurrent introduction of any EHR alerts that could impact ordering (eg, duplicate test warnings) should be simultaneously measured and reported. Studies also need to determine the impact of price display alone compared to price comparison displays (displaying prices for the selected order along with reasonable alternatives). Although price comparison was a component of the intervention in some of the studies in this review, it was not evaluated relative to price display alone. Furthermore, it would be helpful to know if the type of price displayed affects its impact. For instance, if providers are most sensitive to the absolute magnitude of prices, then displaying chargemaster prices may impact ordering more than showing hospital costs. If, however, relative prices are all that providers need, then showing lower numbers, such as Medicare prices or hospital costs, may be sufficient. Finally, it would be reassuring to have additional evidence that price display does not adversely impact patient outcomes.

Although some details need elucidation, the studies synthesized in this review provide valuable data in the current climate of increased emphasis on price transparency. Although substantial attention has been devoted by the academic community, technology start‐ups, private insurers, and even state legislatures to improving price transparency to patients, less focus has been given to physicians, for whom healthcare prices are often just as opaque.[4] The findings from this review suggest that provider price display may be an effective, safe, and acceptable approach to empower physicians to control healthcare spending.

Disclosures: Dr. Silvestri, Dr. Bongiovanni, and Ms. Glover have nothing to disclose. Dr. Gross reports grants from Johnson & Johnson, Medtronic Inc., and 21st Century Oncology during the conduct of this study. In addition, he received payment from Fair Health Inc. and ASTRO outside the submitted work.

As the nation considers how to reduce healthcare costs, hospitalists can play a crucial role in this effort because they control many healthcare services through routine clinical decisions at the point of care. In fact, the government, payers, and the public now look to hospitalists as essential partners for reining in healthcare costs.[1, 2] The role of hospitalists is even more critical as payers, including Medicare, seek to shift reimbursements from volume to value.[1] Medicare's Value‐Based Purchasing program has already tied a percentage of hospital payments to metrics of quality, patient satisfaction, and cost,[1, 3] and Health and Human Services Secretary Sylvia Burwell announced that by the end of 2018, the goal is to have 50% of Medicare payments tied to quality or value through alternative payment models.[4]

Major opportunities for cost savings exist across the care continuum, particularly in postacute and transitional care, and hospitalist groups are leading innovative models that show promise for coordinating care and improving value.[5] Individual hospitalists are also in a unique position to provide high‐value care for their patients through advocating for appropriate care and leading local initiatives to improve value of care.[6, 7, 8] This commentary article aims to provide practicing hospitalists with a framework to incorporate these strategies into their daily work.

DESIGN STRATEGIES TO COORDINATE CARE

As delivery systems undertake the task of population health management, hospitalists will inevitably play a critical role in facilitating coordination between community, acute, and postacute care. During admission, discharge, and the hospitalization itself, standardizing care pathways for common hospital conditions such as pneumonia and cellulitis can be effective in decreasing utilization and improving clinical outcomes.[9, 10] Intermountain Healthcare in Utah has applied evidence‐based protocols to more than 60 clinical processes, re‐engineering roughly 80% of all care that they deliver.[11] These types of care redesigns and standardization promise to provide better, more efficient, and often safer care for more patients. Hospitalists can play important roles in developing and delivering on these pathways.

In addition, hospital physician discontinuity during admissions may lead to increased resource utilization, costs, and lower patient satisfaction.[12] Therefore, ensuring clear handoffs between inpatient providers, as well as with outpatient providers during transitions in care, is a vital component of delivering high‐value care. Of particular importance is the population of patients frequently readmitted to the hospital. Hospitalists are often well acquainted with these patients, and the myriad of psychosocial, economic, and environmental challenges this vulnerable population faces. Although care coordination programs are increasing in prevalence, data on their cost‐effectiveness are mixed, highlighting the need for testing innovations.[13] Certainly, hospitalists can be leaders adopting and documenting the effectiveness of spreading interventions that have been shown to be promising in improving care transitions at discharge, such as the Care Transitions Intervention, Project RED (Re‐Engineered Discharge), or the Transitional Care Model.[14, 15, 16]

The University of Chicago, through funding from the Centers for Medicare and Medicaid Innovation, is testing the use of a single physician who cares for frequently admitted patients both in and out of the hospital, thereby reducing the costs of coordination.[5] This comprehensivist model depends on physicians seeing patients in the hospital and then in a clinic located in or near the hospital for the subset of patients who stand to benefit most from this continuity. This differs from the old model of having primary care providers (PCPs) see inpatients and outpatients because the comprehensivist's patient panel is enriched with only patients who are at high risk for hospitalization, and thus these physicians have a more direct focus on hospital‐related care and higher daily hospitalized patient censuses, whereas PCPs were seeing fewer and fewer of their patients in the hospital on a daily basis. Evidence concerning the effectiveness of this model is expected by 2016. Hospitalists have also ventured out of the hospital into skilled nursing facilities, specializing in long‐term care.[17] These physicians are helping provide care to the roughly 1.6 million residents of US nursing homes.[17, 18] Preliminary evidence suggests increased physician staffing is associated with decreased hospitalization of nursing home residents.[18]

ADVOCATE FOR APPROPRIATE CARE

Hospitalists can advocate for appropriate care through avoiding low‐value services at the point of care, as well as learning and teaching about value.

LEAD LOCAL VALUE INITIATIVES

Hospitalists are ideal leaders of local value initiatives, whether it be through running value‐improvement projects or launching formal high‐value care programs.

Conduct Value‐Improvement Projects

Hospitalists across the country have largely taken the lead on designing value‐improvement pilots, programs, and groups within hospitals. Although value‐improvement projects may be built upon the established structures and techniques for quality improvement, importantly these programs should also include expertise in cost analyses.[8] Furthermore, some traditional quality‐improvement programs have failed to result in actual cost savings[44]; thus, it is not enough to simply rebrand quality improvement with a banner of value. Value‐improvement efforts must overcome the cultural hurdle of more care as better care, as well as pay careful attention to the diplomacy required with value improvement, because reducing costs may result in decreased revenue for certain departments or even decreases in individuals' wages.

One framework that we have used to guide value‐improvement project design is COST: culture, oversight accountability, system support, and training.[45] This approach leverages principles from implementation science to ensure that value‐improvement projects successfully provide multipronged tactics for overcoming the many barriers to high‐value care delivery. Figure 1 includes a worksheet for individual clinicians or teams to use when initially planning value‐improvement project interventions.[46] The examples in this worksheet come from a successful project at the University of California, San Francisco aimed at improving blood utilization stewardship by supporting adherence to a restrictive transfusion strategy. To address culture, a hospital‐wide campaign was led by physician peer champions to raise awareness about appropriate transfusion practices. This included posters that featured prominent local physician leaders displaying their support for the program. Oversight was provided through regular audit and feedback. Each month the number of patients on the medicine service who received transfusion with a pretransfusion hemoglobin above 8 grams per deciliter was shared at a faculty lunch meeting and shown on a graph included in the quality newsletter that was widely distributed in the hospital. The ordering system in the electronic medical record was eventually modified to include the patient's pretransfusion hemoglobin level at time of transfusion order and to provide default options and advice based on whether or not guidelines would generally recommend transfusion. Hospitalists and resident physicians were trained through multiple lectures and informal teaching settings about the rationale behind the changes and the evidence that supported a restrictive transfusion strategy.

Figure 1

CONCLUSIONS

Hospitalists are now faced with a responsibility to reduce financial harm and provide high‐value care. To achieve this goal, hospitalist groups are developing innovative models for care across the continuum from hospital to home, and individual hospitalists can advocate for appropriate care and lead value‐improvement initiatives in hospitals. Through existing knowledge and new frameworks and tools that specifically address value, hospitalists can champion value at the bedside and ensure their patients get the best possible care at lower costs.

Disclosures: Drs. Moriates, Shah, and Arora have received grant funding from the ABIM Foundation, and royalties from McGraw‐Hill for the textbook Understanding Value‐Based Healthcare. The authors report no conflicts of interest.

As the nation considers how to reduce healthcare costs, hospitalists can play a crucial role in this effort because they control many healthcare services through routine clinical decisions at the point of care. In fact, the government, payers, and the public now look to hospitalists as essential partners for reining in healthcare costs.[1, 2] The role of hospitalists is even more critical as payers, including Medicare, seek to shift reimbursements from volume to value.[1] Medicare's Value‐Based Purchasing program has already tied a percentage of hospital payments to metrics of quality, patient satisfaction, and cost,[1, 3] and Health and Human Services Secretary Sylvia Burwell announced that by the end of 2018, the goal is to have 50% of Medicare payments tied to quality or value through alternative payment models.[4]

Major opportunities for cost savings exist across the care continuum, particularly in postacute and transitional care, and hospitalist groups are leading innovative models that show promise for coordinating care and improving value.[5] Individual hospitalists are also in a unique position to provide high‐value care for their patients through advocating for appropriate care and leading local initiatives to improve value of care.[6, 7, 8] This commentary article aims to provide practicing hospitalists with a framework to incorporate these strategies into their daily work.

DESIGN STRATEGIES TO COORDINATE CARE

As delivery systems undertake the task of population health management, hospitalists will inevitably play a critical role in facilitating coordination between community, acute, and postacute care. During admission, discharge, and the hospitalization itself, standardizing care pathways for common hospital conditions such as pneumonia and cellulitis can be effective in decreasing utilization and improving clinical outcomes.[9, 10] Intermountain Healthcare in Utah has applied evidence‐based protocols to more than 60 clinical processes, re‐engineering roughly 80% of all care that they deliver.[11] These types of care redesigns and standardization promise to provide better, more efficient, and often safer care for more patients. Hospitalists can play important roles in developing and delivering on these pathways.

In addition, hospital physician discontinuity during admissions may lead to increased resource utilization, costs, and lower patient satisfaction.[12] Therefore, ensuring clear handoffs between inpatient providers, as well as with outpatient providers during transitions in care, is a vital component of delivering high‐value care. Of particular importance is the population of patients frequently readmitted to the hospital. Hospitalists are often well acquainted with these patients, and the myriad of psychosocial, economic, and environmental challenges this vulnerable population faces. Although care coordination programs are increasing in prevalence, data on their cost‐effectiveness are mixed, highlighting the need for testing innovations.[13] Certainly, hospitalists can be leaders adopting and documenting the effectiveness of spreading interventions that have been shown to be promising in improving care transitions at discharge, such as the Care Transitions Intervention, Project RED (Re‐Engineered Discharge), or the Transitional Care Model.[14, 15, 16]

The University of Chicago, through funding from the Centers for Medicare and Medicaid Innovation, is testing the use of a single physician who cares for frequently admitted patients both in and out of the hospital, thereby reducing the costs of coordination.[5] This comprehensivist model depends on physicians seeing patients in the hospital and then in a clinic located in or near the hospital for the subset of patients who stand to benefit most from this continuity. This differs from the old model of having primary care providers (PCPs) see inpatients and outpatients because the comprehensivist's patient panel is enriched with only patients who are at high risk for hospitalization, and thus these physicians have a more direct focus on hospital‐related care and higher daily hospitalized patient censuses, whereas PCPs were seeing fewer and fewer of their patients in the hospital on a daily basis. Evidence concerning the effectiveness of this model is expected by 2016. Hospitalists have also ventured out of the hospital into skilled nursing facilities, specializing in long‐term care.[17] These physicians are helping provide care to the roughly 1.6 million residents of US nursing homes.[17, 18] Preliminary evidence suggests increased physician staffing is associated with decreased hospitalization of nursing home residents.[18]

ADVOCATE FOR APPROPRIATE CARE

Hospitalists can advocate for appropriate care through avoiding low‐value services at the point of care, as well as learning and teaching about value.

LEAD LOCAL VALUE INITIATIVES

Hospitalists are ideal leaders of local value initiatives, whether it be through running value‐improvement projects or launching formal high‐value care programs.

Conduct Value‐Improvement Projects

Hospitalists across the country have largely taken the lead on designing value‐improvement pilots, programs, and groups within hospitals. Although value‐improvement projects may be built upon the established structures and techniques for quality improvement, importantly these programs should also include expertise in cost analyses.[8] Furthermore, some traditional quality‐improvement programs have failed to result in actual cost savings[44]; thus, it is not enough to simply rebrand quality improvement with a banner of value. Value‐improvement efforts must overcome the cultural hurdle of more care as better care, as well as pay careful attention to the diplomacy required with value improvement, because reducing costs may result in decreased revenue for certain departments or even decreases in individuals' wages.

One framework that we have used to guide value‐improvement project design is COST: culture, oversight accountability, system support, and training.[45] This approach leverages principles from implementation science to ensure that value‐improvement projects successfully provide multipronged tactics for overcoming the many barriers to high‐value care delivery. Figure 1 includes a worksheet for individual clinicians or teams to use when initially planning value‐improvement project interventions.[46] The examples in this worksheet come from a successful project at the University of California, San Francisco aimed at improving blood utilization stewardship by supporting adherence to a restrictive transfusion strategy. To address culture, a hospital‐wide campaign was led by physician peer champions to raise awareness about appropriate transfusion practices. This included posters that featured prominent local physician leaders displaying their support for the program. Oversight was provided through regular audit and feedback. Each month the number of patients on the medicine service who received transfusion with a pretransfusion hemoglobin above 8 grams per deciliter was shared at a faculty lunch meeting and shown on a graph included in the quality newsletter that was widely distributed in the hospital. The ordering system in the electronic medical record was eventually modified to include the patient's pretransfusion hemoglobin level at time of transfusion order and to provide default options and advice based on whether or not guidelines would generally recommend transfusion. Hospitalists and resident physicians were trained through multiple lectures and informal teaching settings about the rationale behind the changes and the evidence that supported a restrictive transfusion strategy.

Figure 1

CONCLUSIONS

Hospitalists are now faced with a responsibility to reduce financial harm and provide high‐value care. To achieve this goal, hospitalist groups are developing innovative models for care across the continuum from hospital to home, and individual hospitalists can advocate for appropriate care and lead value‐improvement initiatives in hospitals. Through existing knowledge and new frameworks and tools that specifically address value, hospitalists can champion value at the bedside and ensure their patients get the best possible care at lower costs.

Disclosures: Drs. Moriates, Shah, and Arora have received grant funding from the ABIM Foundation, and royalties from McGraw‐Hill for the textbook Understanding Value‐Based Healthcare. The authors report no conflicts of interest.

As the nation considers how to reduce healthcare costs, hospitalists can play a crucial role in this effort because they control many healthcare services through routine clinical decisions at the point of care. In fact, the government, payers, and the public now look to hospitalists as essential partners for reining in healthcare costs.[1, 2] The role of hospitalists is even more critical as payers, including Medicare, seek to shift reimbursements from volume to value.[1] Medicare's Value‐Based Purchasing program has already tied a percentage of hospital payments to metrics of quality, patient satisfaction, and cost,[1, 3] and Health and Human Services Secretary Sylvia Burwell announced that by the end of 2018, the goal is to have 50% of Medicare payments tied to quality or value through alternative payment models.[4]

Major opportunities for cost savings exist across the care continuum, particularly in postacute and transitional care, and hospitalist groups are leading innovative models that show promise for coordinating care and improving value.[5] Individual hospitalists are also in a unique position to provide high‐value care for their patients through advocating for appropriate care and leading local initiatives to improve value of care.[6, 7, 8] This commentary article aims to provide practicing hospitalists with a framework to incorporate these strategies into their daily work.

DESIGN STRATEGIES TO COORDINATE CARE

As delivery systems undertake the task of population health management, hospitalists will inevitably play a critical role in facilitating coordination between community, acute, and postacute care. During admission, discharge, and the hospitalization itself, standardizing care pathways for common hospital conditions such as pneumonia and cellulitis can be effective in decreasing utilization and improving clinical outcomes.[9, 10] Intermountain Healthcare in Utah has applied evidence‐based protocols to more than 60 clinical processes, re‐engineering roughly 80% of all care that they deliver.[11] These types of care redesigns and standardization promise to provide better, more efficient, and often safer care for more patients. Hospitalists can play important roles in developing and delivering on these pathways.

In addition, hospital physician discontinuity during admissions may lead to increased resource utilization, costs, and lower patient satisfaction.[12] Therefore, ensuring clear handoffs between inpatient providers, as well as with outpatient providers during transitions in care, is a vital component of delivering high‐value care. Of particular importance is the population of patients frequently readmitted to the hospital. Hospitalists are often well acquainted with these patients, and the myriad of psychosocial, economic, and environmental challenges this vulnerable population faces. Although care coordination programs are increasing in prevalence, data on their cost‐effectiveness are mixed, highlighting the need for testing innovations.[13] Certainly, hospitalists can be leaders adopting and documenting the effectiveness of spreading interventions that have been shown to be promising in improving care transitions at discharge, such as the Care Transitions Intervention, Project RED (Re‐Engineered Discharge), or the Transitional Care Model.[14, 15, 16]

The University of Chicago, through funding from the Centers for Medicare and Medicaid Innovation, is testing the use of a single physician who cares for frequently admitted patients both in and out of the hospital, thereby reducing the costs of coordination.[5] This comprehensivist model depends on physicians seeing patients in the hospital and then in a clinic located in or near the hospital for the subset of patients who stand to benefit most from this continuity. This differs from the old model of having primary care providers (PCPs) see inpatients and outpatients because the comprehensivist's patient panel is enriched with only patients who are at high risk for hospitalization, and thus these physicians have a more direct focus on hospital‐related care and higher daily hospitalized patient censuses, whereas PCPs were seeing fewer and fewer of their patients in the hospital on a daily basis. Evidence concerning the effectiveness of this model is expected by 2016. Hospitalists have also ventured out of the hospital into skilled nursing facilities, specializing in long‐term care.[17] These physicians are helping provide care to the roughly 1.6 million residents of US nursing homes.[17, 18] Preliminary evidence suggests increased physician staffing is associated with decreased hospitalization of nursing home residents.[18]

ADVOCATE FOR APPROPRIATE CARE

Hospitalists can advocate for appropriate care through avoiding low‐value services at the point of care, as well as learning and teaching about value.

LEAD LOCAL VALUE INITIATIVES

Hospitalists are ideal leaders of local value initiatives, whether it be through running value‐improvement projects or launching formal high‐value care programs.

Conduct Value‐Improvement Projects

Hospitalists across the country have largely taken the lead on designing value‐improvement pilots, programs, and groups within hospitals. Although value‐improvement projects may be built upon the established structures and techniques for quality improvement, importantly these programs should also include expertise in cost analyses.[8] Furthermore, some traditional quality‐improvement programs have failed to result in actual cost savings[44]; thus, it is not enough to simply rebrand quality improvement with a banner of value. Value‐improvement efforts must overcome the cultural hurdle of more care as better care, as well as pay careful attention to the diplomacy required with value improvement, because reducing costs may result in decreased revenue for certain departments or even decreases in individuals' wages.

One framework that we have used to guide value‐improvement project design is COST: culture, oversight accountability, system support, and training.[45] This approach leverages principles from implementation science to ensure that value‐improvement projects successfully provide multipronged tactics for overcoming the many barriers to high‐value care delivery. Figure 1 includes a worksheet for individual clinicians or teams to use when initially planning value‐improvement project interventions.[46] The examples in this worksheet come from a successful project at the University of California, San Francisco aimed at improving blood utilization stewardship by supporting adherence to a restrictive transfusion strategy. To address culture, a hospital‐wide campaign was led by physician peer champions to raise awareness about appropriate transfusion practices. This included posters that featured prominent local physician leaders displaying their support for the program. Oversight was provided through regular audit and feedback. Each month the number of patients on the medicine service who received transfusion with a pretransfusion hemoglobin above 8 grams per deciliter was shared at a faculty lunch meeting and shown on a graph included in the quality newsletter that was widely distributed in the hospital. The ordering system in the electronic medical record was eventually modified to include the patient's pretransfusion hemoglobin level at time of transfusion order and to provide default options and advice based on whether or not guidelines would generally recommend transfusion. Hospitalists and resident physicians were trained through multiple lectures and informal teaching settings about the rationale behind the changes and the evidence that supported a restrictive transfusion strategy.

Figure 1

CONCLUSIONS

Hospitalists are now faced with a responsibility to reduce financial harm and provide high‐value care. To achieve this goal, hospitalist groups are developing innovative models for care across the continuum from hospital to home, and individual hospitalists can advocate for appropriate care and lead value‐improvement initiatives in hospitals. Through existing knowledge and new frameworks and tools that specifically address value, hospitalists can champion value at the bedside and ensure their patients get the best possible care at lower costs.

Disclosures: Drs. Moriates, Shah, and Arora have received grant funding from the ABIM Foundation, and royalties from McGraw‐Hill for the textbook Understanding Value‐Based Healthcare. The authors report no conflicts of interest.

Today's medical students will enter practice over the next decade and inherit the escalating costs of the US healthcare system. Approximately 30% of healthcare costs, or $750 billion dollars annually, are spent on unnecessary tests or procedures.[1] High healthcare costs combined with calls to eliminate waste, improve patient safety, and increase quality[2] are driving our healthcare system to evolve from a fee‐based system to a value‐based system. Additionally, many patients are being harmed by overtesting and the stress associated with rising healthcare bills. Financial risk has increasingly shifted to patients in the form of higher deductibles and reduced caps, and medical indebtedness is the number 1 risk for bankruptcy.[3, 4] False positive results of low‐yield diagnostic tests lead to additional testing, anxiety, excess radiation exposure, and unnecessary invasive procedures.[5] To minimize harm to patients, evidence must guide physicians in their ordering behavior. In addition, any care plan a physician develops should be individualized to incorporate patients' values and preferences. Unfortunately, medical students, who are at an impressionable stage in their careers, frequently observe overtesting and unnecessary treatment behaviors in their clinical encounters.[6] Instead, our medical students and trainees must be prepared to deliver patient care that is evidence based, patient centered, and cost conscious. They must become effective stewards of limited healthcare resources.

To help prepare our students for this evolving healthcare paradigm, we created a new tool called SOAP‐V (Subjective‐Objective‐Assessment‐PlanValue), designed to embed discussion of healthcare value into medical student oral presentations and note writing. Students are encouraged to use this tool at the point of care to bring up value concepts with physicians and residents as part of medical decision making. In so doing, we propose that medical students can serve as change agents to shift physician practice at our academic medical centers to focus on healthcare value. This article describes the SOAP‐V tool, contains links to educational materials to help hospitalists and other clinician educators to implement this tool, and provides preliminary findings and reflections.

INNOVATION

SOAP‐V was conceived at the Millennium Conference on Teaching High‐Value Care, which was sponsored by the Beth Israel Deaconess Medical Center Shapiro Institute for Education and Research, the Association of American Medical Colleges, and the American College of Physicians. Educators from several medical schools decided to form a group to specifically consider ways to train medical students and residents in the concept of high‐value care (HVC), which is framed as improving patient outcomes while decreasing patient cost and harm.[7] Our group recognized several challenges in teaching HVC. First, physician practice habits are influenced by the way they are trained,[8] yet faculty who teach those future physicians frequently have not themselves been taught, nor do they consistently practice HVC.[9] Second, we needed to teach students the requisite HVC knowledge, attitudes, and skills, and therefore wanted to provide opportunities to not only learn, but practice, HVC, preferably in authentic patient experiences to optimize their learning.[10, 11] Third, we recognized that adding another teaching task to the already oversubscribed day of an attending might understandably be met with resistance. We envisioned a tool that could be used with minimal or no faculty training, could be attached to authentic patient experiences, and based on LEAN‐Six Sigma principles,[12] would be embedded in the normal workflow. Furthermore, we considered social networking principles, such as those described by Christakis and Fowler that describe how an individual's behavior impacts behaviors of those surrounding them,[13] and hoped to empower medical students to serve as change agents. Medical students could initiate discussions of value concepts at the point of care in a way that challenges a heavily entrenched test‐ordering culture and encourages other members of the team to balance potential benefit with harms and cost. Following the conference, the group held bimonthly phone conferences and subsequently developed the SOAP‐V tool, created teaching materials, and planned a research project on SOAP‐V.

SOAP‐V modifies the traditional SOAP (Subjective‐Objective‐Assessment‐Plan) oral presentation or medical note, to include value (V). It serves as a cognitive forcing function designed to create a pause and promote discussions of HVC during patient delivery. It prompts the student to ask 3 value questions: (1) Before choosing an intervention, have I considered whether the result would change management? (2) Have I incorporated the patient's goals and values, and considered the potential harm of the intervention compared to alternatives? (3) What is the known and potential cost of the intervention, both immediate and downstream? The student gathers information during the patient interview and brings back the information to the team during rounds where management decisions are made.

In the summer of 2014, we launched an institutional review boardapproved, multi‐institutional study to implement SOAP‐V at Penn State College of Medicine, Harvard Medical School, and Case Western Reserve University School of Medicine for third‐year medical students during their internal medicine clerkships. Students in the intervention arm participated in an interactive workshop on SOAP‐V. Authors S.F., S.G., and C.D.P., who serve as clerkship directors in internal medicine, provided student training for each cohort of intervention students at the beginning of each rotation on general medicine inpatient wards. The workshop began with trigger videos that demonstrate pressures encountered by a student on rounds that might lead to overuse.[14] Following a discussion on overuse and methods to avoid overuse, the students were introduced to the SOAP‐V framework, watched a video of a student modeling a SOAP‐V presentation on rounds,[15] and engaged in a SOAP‐V role play. They received a SOAP‐V pocket card as well as a Web link to Healthcare Bluebook[16] to research costs. An outline of the session and associated materials can be found in an online attachment.[17] The students then used the SOAP‐V tool during inpatient rounds. We advised supervising faculty that students might present using a SOAP‐V format, and provided them with a SOAP‐V card, but we did not provide faculty development on SOAP‐V. Students participating in the control arm did not receive training specific to SOAP‐V.

Students in intervention and control arms at each school were surveyed on their attitudes toward HVC at the beginning of the clerkship year and then again at the completion of the medicine clerkship via a 19‐item questionnaire soliciting perceptions and self‐reported practices in HVC. Intervention arm students received biweekly e‐mail links that allowed them to anonymously document their use of SOAP‐V, as well as an end‐of‐clerkship open‐ended question about the usefulness of SOAP‐V. We analyzed questionnaire results using McNemar's test for paired data.

PRELIMINARY FINDINGS

The preintervention attitudinal survey (n = 226) demonstrated that although 90% of medical students agreed on the importance of considering costs of treatments, only 50% felt comfortable bringing up cost considerations with their team, and 50% considered costs to the healthcare system in clinical decisions. An interim analysis of the available data at 6 months (response rate approximately 50% across sites) showed that students in the intervention arm reported increased agreement with the phrases, I have the power to address the economic healthcare crisis (pre‐37%, post‐65%, P = 0.046); I would be comfortable initiating a discussion about unnecessary tests or treatments with my team, (pre‐46%, post‐85%, P = 0.027); and In my clinical decisions, I consider the potential costs to the healthcare system (pre‐41%, post‐60%, P = 0.023) compared to control arm students, who showed no significant differences pre‐ versus postrotation in these 3 domains (Figure 1).

Figure 1

To date, biweekly surveys and direct observation of rounds have verified student use of SOAP‐V. Student comments have included: Allowed me the ability to raise important issues with the team while feeling like I was helping my patients and the healthcare system. A great principle that I used almost daily. Great to implement this at such a young stage in my med career. Broadened my perspective on the role of a physician.

SOAP‐V has inspired some of our medical students to consider value in healthcare more closely. In a notable example, a SOAP‐Vtrained student admitted a young man with lymphadenopathy, pulmonary infiltrates, and weight loss who underwent an extensive and costly workup including liver biopsy, bronchoscopy, and multiple computed tomography and positron emission tomography scans and was eventually diagnosed with sarcoidosis. The SOAP‐Vtrained student reviewed the patient's workup, estimated that the team spent more than $6000 to make the diagnosis, and recommended a more cost‐effective approach.

Common barriers experienced by the pilot sites included time constraints limiting discussion of value, variability in perceived receptivity depending on team leadership, and student confidence in initiating this dialogue. Solutions included underscoring the notion that value discussions can be brief, may be appropriately initiated by any member of the team, and may have an effect on choice of management and/or patient preference issues that can make medical care more efficient and effective. Resident and faculty physicians were made aware of the intervention, and encouraged to support students in using the SOAP‐V tool.

CONCLUSION

SOAP‐V was successfully implemented within the inpatient internal medicine clerkship at 3 academic institutions. Our preliminary results demonstrate that students can use this framework to apply considerations of high‐value, cost‐conscious care in their medical decision making and to promote discussion of these concepts during rounds with their inpatient teams. Students in the intervention arm report greater comfort discussing unnecessary tests and treatments with their team and a greater likelihood to consider potential costs to the healthcare system. Additionally, these students commented that the SOAP‐V framework broadened their perspective on their role as a physician in curbing costs, and that they felt more empowered to address the economic healthcare crisis. The next phase of our project will involve conducting end‐of‐year surveys to evaluate whether SOAP‐V has a persistent impact on the frequency and quality of value discussions on rounds, as well as students' attitudes about cost consciousness. We will also gauge whether resident and faculty attitudes about HVC have changed as a result of the intervention.

Our SOAP‐V student training was provided in a 1‐hour session. We believe that the ease of training and the simplicity of the SOAP‐V framework permit SOAP‐V to be easily transferred for use by residents, medical students in other clerkships, and other healthcare learners. Additional research is needed to demonstrate this expanded use and prove sustainability. An additional important question is whether use of SOAP‐V by students and residents results in reductions in unnecessary costs. Future educational efforts will include embedding the SOAP‐V tool in other clerkships and promoting the SOAP‐V tool within corresponding residencies in both hospital and outpatient clinic settings and analyzing potential reductions in wasteful spending.

It is generally conceived that medical students learn the information they are taught, and are impacted by the culture in which they reside; multiple studies bear this out.[18, 19] However, students may also be change agents. Our students will inherit the healthcare systems of the future. We must empower them to change the status quo. There can be tremendous utility in employing such a bottom up approach to process improvement. What a student discusses today may spark the resident (or faculty) to consider in their own workflow tomorrow. In this way, we envision that the SOAP‐V is a tool by which ideas concerning HVC can be generated and shared at the point of care. It is our hope that this straightforward intervention is one that may slowly change the culture and perhaps eventually the practice patterns of our academic medical centers.

Today's medical students will enter practice over the next decade and inherit the escalating costs of the US healthcare system. Approximately 30% of healthcare costs, or $750 billion dollars annually, are spent on unnecessary tests or procedures.[1] High healthcare costs combined with calls to eliminate waste, improve patient safety, and increase quality[2] are driving our healthcare system to evolve from a fee‐based system to a value‐based system. Additionally, many patients are being harmed by overtesting and the stress associated with rising healthcare bills. Financial risk has increasingly shifted to patients in the form of higher deductibles and reduced caps, and medical indebtedness is the number 1 risk for bankruptcy.[3, 4] False positive results of low‐yield diagnostic tests lead to additional testing, anxiety, excess radiation exposure, and unnecessary invasive procedures.[5] To minimize harm to patients, evidence must guide physicians in their ordering behavior. In addition, any care plan a physician develops should be individualized to incorporate patients' values and preferences. Unfortunately, medical students, who are at an impressionable stage in their careers, frequently observe overtesting and unnecessary treatment behaviors in their clinical encounters.[6] Instead, our medical students and trainees must be prepared to deliver patient care that is evidence based, patient centered, and cost conscious. They must become effective stewards of limited healthcare resources.

To help prepare our students for this evolving healthcare paradigm, we created a new tool called SOAP‐V (Subjective‐Objective‐Assessment‐PlanValue), designed to embed discussion of healthcare value into medical student oral presentations and note writing. Students are encouraged to use this tool at the point of care to bring up value concepts with physicians and residents as part of medical decision making. In so doing, we propose that medical students can serve as change agents to shift physician practice at our academic medical centers to focus on healthcare value. This article describes the SOAP‐V tool, contains links to educational materials to help hospitalists and other clinician educators to implement this tool, and provides preliminary findings and reflections.

INNOVATION

SOAP‐V was conceived at the Millennium Conference on Teaching High‐Value Care, which was sponsored by the Beth Israel Deaconess Medical Center Shapiro Institute for Education and Research, the Association of American Medical Colleges, and the American College of Physicians. Educators from several medical schools decided to form a group to specifically consider ways to train medical students and residents in the concept of high‐value care (HVC), which is framed as improving patient outcomes while decreasing patient cost and harm.[7] Our group recognized several challenges in teaching HVC. First, physician practice habits are influenced by the way they are trained,[8] yet faculty who teach those future physicians frequently have not themselves been taught, nor do they consistently practice HVC.[9] Second, we needed to teach students the requisite HVC knowledge, attitudes, and skills, and therefore wanted to provide opportunities to not only learn, but practice, HVC, preferably in authentic patient experiences to optimize their learning.[10, 11] Third, we recognized that adding another teaching task to the already oversubscribed day of an attending might understandably be met with resistance. We envisioned a tool that could be used with minimal or no faculty training, could be attached to authentic patient experiences, and based on LEAN‐Six Sigma principles,[12] would be embedded in the normal workflow. Furthermore, we considered social networking principles, such as those described by Christakis and Fowler that describe how an individual's behavior impacts behaviors of those surrounding them,[13] and hoped to empower medical students to serve as change agents. Medical students could initiate discussions of value concepts at the point of care in a way that challenges a heavily entrenched test‐ordering culture and encourages other members of the team to balance potential benefit with harms and cost. Following the conference, the group held bimonthly phone conferences and subsequently developed the SOAP‐V tool, created teaching materials, and planned a research project on SOAP‐V.

SOAP‐V modifies the traditional SOAP (Subjective‐Objective‐Assessment‐Plan) oral presentation or medical note, to include value (V). It serves as a cognitive forcing function designed to create a pause and promote discussions of HVC during patient delivery. It prompts the student to ask 3 value questions: (1) Before choosing an intervention, have I considered whether the result would change management? (2) Have I incorporated the patient's goals and values, and considered the potential harm of the intervention compared to alternatives? (3) What is the known and potential cost of the intervention, both immediate and downstream? The student gathers information during the patient interview and brings back the information to the team during rounds where management decisions are made.

In the summer of 2014, we launched an institutional review boardapproved, multi‐institutional study to implement SOAP‐V at Penn State College of Medicine, Harvard Medical School, and Case Western Reserve University School of Medicine for third‐year medical students during their internal medicine clerkships. Students in the intervention arm participated in an interactive workshop on SOAP‐V. Authors S.F., S.G., and C.D.P., who serve as clerkship directors in internal medicine, provided student training for each cohort of intervention students at the beginning of each rotation on general medicine inpatient wards. The workshop began with trigger videos that demonstrate pressures encountered by a student on rounds that might lead to overuse.[14] Following a discussion on overuse and methods to avoid overuse, the students were introduced to the SOAP‐V framework, watched a video of a student modeling a SOAP‐V presentation on rounds,[15] and engaged in a SOAP‐V role play. They received a SOAP‐V pocket card as well as a Web link to Healthcare Bluebook[16] to research costs. An outline of the session and associated materials can be found in an online attachment.[17] The students then used the SOAP‐V tool during inpatient rounds. We advised supervising faculty that students might present using a SOAP‐V format, and provided them with a SOAP‐V card, but we did not provide faculty development on SOAP‐V. Students participating in the control arm did not receive training specific to SOAP‐V.

Students in intervention and control arms at each school were surveyed on their attitudes toward HVC at the beginning of the clerkship year and then again at the completion of the medicine clerkship via a 19‐item questionnaire soliciting perceptions and self‐reported practices in HVC. Intervention arm students received biweekly e‐mail links that allowed them to anonymously document their use of SOAP‐V, as well as an end‐of‐clerkship open‐ended question about the usefulness of SOAP‐V. We analyzed questionnaire results using McNemar's test for paired data.

PRELIMINARY FINDINGS

The preintervention attitudinal survey (n = 226) demonstrated that although 90% of medical students agreed on the importance of considering costs of treatments, only 50% felt comfortable bringing up cost considerations with their team, and 50% considered costs to the healthcare system in clinical decisions. An interim analysis of the available data at 6 months (response rate approximately 50% across sites) showed that students in the intervention arm reported increased agreement with the phrases, I have the power to address the economic healthcare crisis (pre‐37%, post‐65%, P = 0.046); I would be comfortable initiating a discussion about unnecessary tests or treatments with my team, (pre‐46%, post‐85%, P = 0.027); and In my clinical decisions, I consider the potential costs to the healthcare system (pre‐41%, post‐60%, P = 0.023) compared to control arm students, who showed no significant differences pre‐ versus postrotation in these 3 domains (Figure 1).

Figure 1

To date, biweekly surveys and direct observation of rounds have verified student use of SOAP‐V. Student comments have included: Allowed me the ability to raise important issues with the team while feeling like I was helping my patients and the healthcare system. A great principle that I used almost daily. Great to implement this at such a young stage in my med career. Broadened my perspective on the role of a physician.

SOAP‐V has inspired some of our medical students to consider value in healthcare more closely. In a notable example, a SOAP‐Vtrained student admitted a young man with lymphadenopathy, pulmonary infiltrates, and weight loss who underwent an extensive and costly workup including liver biopsy, bronchoscopy, and multiple computed tomography and positron emission tomography scans and was eventually diagnosed with sarcoidosis. The SOAP‐Vtrained student reviewed the patient's workup, estimated that the team spent more than $6000 to make the diagnosis, and recommended a more cost‐effective approach.

Common barriers experienced by the pilot sites included time constraints limiting discussion of value, variability in perceived receptivity depending on team leadership, and student confidence in initiating this dialogue. Solutions included underscoring the notion that value discussions can be brief, may be appropriately initiated by any member of the team, and may have an effect on choice of management and/or patient preference issues that can make medical care more efficient and effective. Resident and faculty physicians were made aware of the intervention, and encouraged to support students in using the SOAP‐V tool.

CONCLUSION

SOAP‐V was successfully implemented within the inpatient internal medicine clerkship at 3 academic institutions. Our preliminary results demonstrate that students can use this framework to apply considerations of high‐value, cost‐conscious care in their medical decision making and to promote discussion of these concepts during rounds with their inpatient teams. Students in the intervention arm report greater comfort discussing unnecessary tests and treatments with their team and a greater likelihood to consider potential costs to the healthcare system. Additionally, these students commented that the SOAP‐V framework broadened their perspective on their role as a physician in curbing costs, and that they felt more empowered to address the economic healthcare crisis. The next phase of our project will involve conducting end‐of‐year surveys to evaluate whether SOAP‐V has a persistent impact on the frequency and quality of value discussions on rounds, as well as students' attitudes about cost consciousness. We will also gauge whether resident and faculty attitudes about HVC have changed as a result of the intervention.

Our SOAP‐V student training was provided in a 1‐hour session. We believe that the ease of training and the simplicity of the SOAP‐V framework permit SOAP‐V to be easily transferred for use by residents, medical students in other clerkships, and other healthcare learners. Additional research is needed to demonstrate this expanded use and prove sustainability. An additional important question is whether use of SOAP‐V by students and residents results in reductions in unnecessary costs. Future educational efforts will include embedding the SOAP‐V tool in other clerkships and promoting the SOAP‐V tool within corresponding residencies in both hospital and outpatient clinic settings and analyzing potential reductions in wasteful spending.

It is generally conceived that medical students learn the information they are taught, and are impacted by the culture in which they reside; multiple studies bear this out.[18, 19] However, students may also be change agents. Our students will inherit the healthcare systems of the future. We must empower them to change the status quo. There can be tremendous utility in employing such a bottom up approach to process improvement. What a student discusses today may spark the resident (or faculty) to consider in their own workflow tomorrow. In this way, we envision that the SOAP‐V is a tool by which ideas concerning HVC can be generated and shared at the point of care. It is our hope that this straightforward intervention is one that may slowly change the culture and perhaps eventually the practice patterns of our academic medical centers.

Today's medical students will enter practice over the next decade and inherit the escalating costs of the US healthcare system. Approximately 30% of healthcare costs, or $750 billion dollars annually, are spent on unnecessary tests or procedures.[1] High healthcare costs combined with calls to eliminate waste, improve patient safety, and increase quality[2] are driving our healthcare system to evolve from a fee‐based system to a value‐based system. Additionally, many patients are being harmed by overtesting and the stress associated with rising healthcare bills. Financial risk has increasingly shifted to patients in the form of higher deductibles and reduced caps, and medical indebtedness is the number 1 risk for bankruptcy.[3, 4] False positive results of low‐yield diagnostic tests lead to additional testing, anxiety, excess radiation exposure, and unnecessary invasive procedures.[5] To minimize harm to patients, evidence must guide physicians in their ordering behavior. In addition, any care plan a physician develops should be individualized to incorporate patients' values and preferences. Unfortunately, medical students, who are at an impressionable stage in their careers, frequently observe overtesting and unnecessary treatment behaviors in their clinical encounters.[6] Instead, our medical students and trainees must be prepared to deliver patient care that is evidence based, patient centered, and cost conscious. They must become effective stewards of limited healthcare resources.

To help prepare our students for this evolving healthcare paradigm, we created a new tool called SOAP‐V (Subjective‐Objective‐Assessment‐PlanValue), designed to embed discussion of healthcare value into medical student oral presentations and note writing. Students are encouraged to use this tool at the point of care to bring up value concepts with physicians and residents as part of medical decision making. In so doing, we propose that medical students can serve as change agents to shift physician practice at our academic medical centers to focus on healthcare value. This article describes the SOAP‐V tool, contains links to educational materials to help hospitalists and other clinician educators to implement this tool, and provides preliminary findings and reflections.

INNOVATION

SOAP‐V was conceived at the Millennium Conference on Teaching High‐Value Care, which was sponsored by the Beth Israel Deaconess Medical Center Shapiro Institute for Education and Research, the Association of American Medical Colleges, and the American College of Physicians. Educators from several medical schools decided to form a group to specifically consider ways to train medical students and residents in the concept of high‐value care (HVC), which is framed as improving patient outcomes while decreasing patient cost and harm.[7] Our group recognized several challenges in teaching HVC. First, physician practice habits are influenced by the way they are trained,[8] yet faculty who teach those future physicians frequently have not themselves been taught, nor do they consistently practice HVC.[9] Second, we needed to teach students the requisite HVC knowledge, attitudes, and skills, and therefore wanted to provide opportunities to not only learn, but practice, HVC, preferably in authentic patient experiences to optimize their learning.[10, 11] Third, we recognized that adding another teaching task to the already oversubscribed day of an attending might understandably be met with resistance. We envisioned a tool that could be used with minimal or no faculty training, could be attached to authentic patient experiences, and based on LEAN‐Six Sigma principles,[12] would be embedded in the normal workflow. Furthermore, we considered social networking principles, such as those described by Christakis and Fowler that describe how an individual's behavior impacts behaviors of those surrounding them,[13] and hoped to empower medical students to serve as change agents. Medical students could initiate discussions of value concepts at the point of care in a way that challenges a heavily entrenched test‐ordering culture and encourages other members of the team to balance potential benefit with harms and cost. Following the conference, the group held bimonthly phone conferences and subsequently developed the SOAP‐V tool, created teaching materials, and planned a research project on SOAP‐V.

SOAP‐V modifies the traditional SOAP (Subjective‐Objective‐Assessment‐Plan) oral presentation or medical note, to include value (V). It serves as a cognitive forcing function designed to create a pause and promote discussions of HVC during patient delivery. It prompts the student to ask 3 value questions: (1) Before choosing an intervention, have I considered whether the result would change management? (2) Have I incorporated the patient's goals and values, and considered the potential harm of the intervention compared to alternatives? (3) What is the known and potential cost of the intervention, both immediate and downstream? The student gathers information during the patient interview and brings back the information to the team during rounds where management decisions are made.

In the summer of 2014, we launched an institutional review boardapproved, multi‐institutional study to implement SOAP‐V at Penn State College of Medicine, Harvard Medical School, and Case Western Reserve University School of Medicine for third‐year medical students during their internal medicine clerkships. Students in the intervention arm participated in an interactive workshop on SOAP‐V. Authors S.F., S.G., and C.D.P., who serve as clerkship directors in internal medicine, provided student training for each cohort of intervention students at the beginning of each rotation on general medicine inpatient wards. The workshop began with trigger videos that demonstrate pressures encountered by a student on rounds that might lead to overuse.[14] Following a discussion on overuse and methods to avoid overuse, the students were introduced to the SOAP‐V framework, watched a video of a student modeling a SOAP‐V presentation on rounds,[15] and engaged in a SOAP‐V role play. They received a SOAP‐V pocket card as well as a Web link to Healthcare Bluebook[16] to research costs. An outline of the session and associated materials can be found in an online attachment.[17] The students then used the SOAP‐V tool during inpatient rounds. We advised supervising faculty that students might present using a SOAP‐V format, and provided them with a SOAP‐V card, but we did not provide faculty development on SOAP‐V. Students participating in the control arm did not receive training specific to SOAP‐V.

Students in intervention and control arms at each school were surveyed on their attitudes toward HVC at the beginning of the clerkship year and then again at the completion of the medicine clerkship via a 19‐item questionnaire soliciting perceptions and self‐reported practices in HVC. Intervention arm students received biweekly e‐mail links that allowed them to anonymously document their use of SOAP‐V, as well as an end‐of‐clerkship open‐ended question about the usefulness of SOAP‐V. We analyzed questionnaire results using McNemar's test for paired data.

PRELIMINARY FINDINGS

The preintervention attitudinal survey (n = 226) demonstrated that although 90% of medical students agreed on the importance of considering costs of treatments, only 50% felt comfortable bringing up cost considerations with their team, and 50% considered costs to the healthcare system in clinical decisions. An interim analysis of the available data at 6 months (response rate approximately 50% across sites) showed that students in the intervention arm reported increased agreement with the phrases, I have the power to address the economic healthcare crisis (pre‐37%, post‐65%, P = 0.046); I would be comfortable initiating a discussion about unnecessary tests or treatments with my team, (pre‐46%, post‐85%, P = 0.027); and In my clinical decisions, I consider the potential costs to the healthcare system (pre‐41%, post‐60%, P = 0.023) compared to control arm students, who showed no significant differences pre‐ versus postrotation in these 3 domains (Figure 1).

Figure 1

To date, biweekly surveys and direct observation of rounds have verified student use of SOAP‐V. Student comments have included: Allowed me the ability to raise important issues with the team while feeling like I was helping my patients and the healthcare system. A great principle that I used almost daily. Great to implement this at such a young stage in my med career. Broadened my perspective on the role of a physician.

SOAP‐V has inspired some of our medical students to consider value in healthcare more closely. In a notable example, a SOAP‐Vtrained student admitted a young man with lymphadenopathy, pulmonary infiltrates, and weight loss who underwent an extensive and costly workup including liver biopsy, bronchoscopy, and multiple computed tomography and positron emission tomography scans and was eventually diagnosed with sarcoidosis. The SOAP‐Vtrained student reviewed the patient's workup, estimated that the team spent more than $6000 to make the diagnosis, and recommended a more cost‐effective approach.

Common barriers experienced by the pilot sites included time constraints limiting discussion of value, variability in perceived receptivity depending on team leadership, and student confidence in initiating this dialogue. Solutions included underscoring the notion that value discussions can be brief, may be appropriately initiated by any member of the team, and may have an effect on choice of management and/or patient preference issues that can make medical care more efficient and effective. Resident and faculty physicians were made aware of the intervention, and encouraged to support students in using the SOAP‐V tool.

CONCLUSION

SOAP‐V was successfully implemented within the inpatient internal medicine clerkship at 3 academic institutions. Our preliminary results demonstrate that students can use this framework to apply considerations of high‐value, cost‐conscious care in their medical decision making and to promote discussion of these concepts during rounds with their inpatient teams. Students in the intervention arm report greater comfort discussing unnecessary tests and treatments with their team and a greater likelihood to consider potential costs to the healthcare system. Additionally, these students commented that the SOAP‐V framework broadened their perspective on their role as a physician in curbing costs, and that they felt more empowered to address the economic healthcare crisis. The next phase of our project will involve conducting end‐of‐year surveys to evaluate whether SOAP‐V has a persistent impact on the frequency and quality of value discussions on rounds, as well as students' attitudes about cost consciousness. We will also gauge whether resident and faculty attitudes about HVC have changed as a result of the intervention.

Our SOAP‐V student training was provided in a 1‐hour session. We believe that the ease of training and the simplicity of the SOAP‐V framework permit SOAP‐V to be easily transferred for use by residents, medical students in other clerkships, and other healthcare learners. Additional research is needed to demonstrate this expanded use and prove sustainability. An additional important question is whether use of SOAP‐V by students and residents results in reductions in unnecessary costs. Future educational efforts will include embedding the SOAP‐V tool in other clerkships and promoting the SOAP‐V tool within corresponding residencies in both hospital and outpatient clinic settings and analyzing potential reductions in wasteful spending.

It is generally conceived that medical students learn the information they are taught, and are impacted by the culture in which they reside; multiple studies bear this out.[18, 19] However, students may also be change agents. Our students will inherit the healthcare systems of the future. We must empower them to change the status quo. There can be tremendous utility in employing such a bottom up approach to process improvement. What a student discusses today may spark the resident (or faculty) to consider in their own workflow tomorrow. In this way, we envision that the SOAP‐V is a tool by which ideas concerning HVC can be generated and shared at the point of care. It is our hope that this straightforward intervention is one that may slowly change the culture and perhaps eventually the practice patterns of our academic medical centers.

Inpatient hospital services are a major component of total US civilian noninstitutionalized healthcare expenses, accounting for 29.3% of spending in 2009[1] when the average cost per stay was $9700.[2] Telemetry monitoring, a widely used resource for the identification of life‐threatening arrhythmias, contributes to these costs. In 1998, Sivaram et al. estimated the cost per patient at $683; in 2010, Ivonye et al. published the cost difference between a telemetry bed and a nonmonitored bed in their inner‐city public teaching facility reached $800.[3, 4]

In 1991, the American College of Cardiology published guidelines for telemetry use, which were later revised by the American Heart Association in 2004.[5, 6] Notably, the guidelines are based on expert opinion and on research data in electrocardiography.[7] The guidelines divide patients into 3 classes based on clinical condition: recommending telemetry monitoring for almost all class I patients, stating possible benefit in class II patients, and discouraging cardiac monitoring for the low‐risk class III patients.[5, 6] The Choosing Wisely campaign, an initiative of the American Board of Internal Medicine and the Society of Hospital Medicine, highlights telemetry monitoring as 1 of the top 5 interventions that physicians and patients should question when determining tests and procedures.[8] Choosing Wisely suggests using a protocol to govern continuation of telemetry outside of the intensive care unit (ICU), as inappropriate monitoring increases care costs and may result in patient harm.[8] The Joint Commission 2014 National Patient Safety Goals notes that numerous alarm signals and the resulting noise and displayed information tends to desensitize staff and cause them to miss or ignore alarm signals or even disable them.[9]

Few studies have examined implementation methods for improved telemetry bed utilization. One study evaluated the impact of a multispecialty telemetry policy with enforcement by an outside cardiologist and nurse team, noting improved cardiac monitoring bed utilization and decreased academic hospital closure, which previously resulted in inability to accept new patients or procedure cancellation.[10] Another study provided an orientation handout discussed by the chief resident and telemetry indication reviews during multidisciplinary rounds 3 times a week.[11]

Our study is one the first to demonstrate a model for a hospitalist‐led approach to guide appropriate telemetry use. We investigated the impact of a multipronged approach to guide telemetry usage: (1) a hospitalist‐led, daily review of bed utilization during attending rounds, (2) a hospitalist attending‐driven, trainee‐focused education module on telemetry utilization, (3) quarterly feedback on telemetry bed utilization rates, and (4) financial incentives. We analyzed pre‐ and post‐evaluation results from the education module to measure impact on knowledge, skills, and attitudes. Additionally, we evaluated the effect of the intervention on length of stay (LOS) and bed utilization costs, while monitoring case mix index (CMI) and overall mortality.

METHODS

RESULTS

DISCUSSION

Our study is one of the first to our knowledge to demonstrate reductions in telemetry LOS by a hospitalist intervention for telemetry utilization. Others[10, 11] have studied the impact of an orientation handout by chief residents or a multispecialty telemetry policy with enforcement by an outside cardiologist and nurse team. Dressler et al. later sustained a 70% reduction in telemetry use without adversely affecting patient safety, as assessed through numbers of rapid response activations, codes, and deaths, through integrating the AHA guidelines into their electronic ordering system.[12] However, our study has the advantage of the primary team, who knows the patient and clinical scenario best, driving the change during attending rounds. In an era where cost consciousness intersects the practice of medicine, any intervention in patient care that demonstrates cost savings without an adverse impact on patient care and resource utilization must be emphasized. This is particularly important in academic institutions, where residents and medical students are learning to integrate the principles of patient safety and quality improvement into their clinical practice.[13] We actually showed sustained telemetry LOS reductions into the extension period after our intervention. We believe this may be due to telemetry triage being integrated into our attending and resident rounding practices. Future work should include integration of telemetry triage into clinical decision support in the electronic medical record and multidisciplinary rounds to disseminate telemetry triage hospital‐wide in both the academic and community settings.

Our study also revealed that nearly half of participants were not aware of the criteria for appropriate utilization of telemetry before our intervention; in the preintervention period, there were many anecdotal and objective findings of inappropriate utilization of telemetry as well as prolonged continuation beyond the clinical needs in both the hospitalist and nonhospitalist group. For the hospitalist group (ie, the group receiving guideline‐based education on appropriate indications for telemetry utilization), there was an assessment of both appropriate usage and timely discontinuation of telemetry in the postintervention period, which we attribute in large part to adherence to the education provided to this group.

We were able to show increased knowledge of cost‐saving actions among trainees with our educational module. We believe it is imperative to educate our providers (physicians, nurses, case managers, and students within these disciplines) on the appropriate indications for telemetry use, not only to help with cost savings and resource availability (ie, allowing telemetry beds to be available for patients who need them most), but also to instill consistent expectations among our patients. For the hospitalist group (ie, the group receiving guideline‐based education on appropriate indications for telemetry utilization), there was an assessment of both appropriate usage and timely discontinuation of telemetry in the postintervention period, which we attribute in large part to adherence to the education provided to this group.

Additionally, we feel it is important to consider the impacts of inappropriate use of telemetry from a patient's perspective: it is physically restrictive/emnconvenient, alarms are disruptive, it can be a barrier for other treatments such as physical therapy, it may increase the time it takes for imaging studies, a nurse may be required to accompany patients on telemetry, and poses additional costs to their medical bill.

We believe our success is due to several strategies. First, at the start of the fiscal year when quality improvement metrics are established, this particular metric (improving the appropriate utilization and timely discontinuation of telemetry) was deemed important by all hospitalists, engendering group buy‐in prior to the intervention. Our hospitalists received a detailed and interactive tutorial session in person at the beginning of the study. This tutorial provided the hospitalists with a comprehensive understanding of the appropriate (and inappropriate) indications for telemetry monitoring, hence facilitating guideline‐directed utilization. Email reminders and the tutorial tool were provided each time a hospitalist attended on the wards, and hospitalists received a small financial incentive to comply with appropriate telemetry utilization.

Our study has several strengths. First, the time frame of our study was long enough (8 months) to allow consistent trends to emerge and to optimize exposure of housestaff and medical students to this quality‐improvement initiative. Second, our cost savings came from 2 factors, direct reduction of inappropriate telemetry use and reduction in length of stay, highlighting the dual impact of appropriate telemetry utilization on cost. The overall reductions in telemetry utilization for the intervention group were a result of both reductions in initial placement on telemetry for patients who did not meet criteria for such monitoring as well as timely discontinuation of telemetry during the patient's hospitalization. Third, our study demonstrates that physicians can be effective in driving appropriate telemetry usage by participating in the clinical decision making regarding necessity and educating providers, trainees/students, and patients on appropriate indications. Finally, we show sustainment of our intervention in the extension period, suggesting telemetry triage integration into rounding practice.

Our study has limitations as well. First, our sample size is relatively small at a single academic center. Second, due to complexities in our faculty scheduling, we were unable to completely randomize patients to a hospitalist versus nonhospitalist team. However, we believe that despite the inability to randomize, our study does show the benefit of a hospitalist attending to reduce telemetry LOS given there was no change in nonhospitalist telemetry LOS despite all of the other hospital‐wide interventions (multidisciplinary rounds, similar housestaff). Third, our study was limited in that the CMI was used as a proxy for patient complexity, and the mortality index was used as the overall marker of safety. Further studies should monitor frequency and outcomes of arrhythmic events of patients transferred from telemetry monitoring to medicalsurgical beds. Finally, as the intervention was multipronged, we are unable to determine which component led to the reductions in telemetry utilization. Each component, however, remains easily transferrable to outside institutions. We demonstrated both a reduction in initiation of telemetry as well as timely discontinuation; however, due to the complexity in capturing this accurately, we were unable to numerically quantify these individual outcomes.

Additionally, there were approximately 10 nonhospitalist attendings who also staffed the wards during the intervention time period of our study; these attendings did not undergo the telemetry tutorial/orientation. This difference, along with the Hawthorne effect for the hospitalist attendings, also likely contributed to the difference in outcomes between the 2 attending cohorts in the intervention period.

CONCLUSIONS

Our results demonstrate that a multipronged hospitalist‐driven intervention to improve appropriate use of telemetry reduces telemetry LOS and cost. Hence, we believe that targeted, education‐driven interventions with monitoring of progress can have demonstrable impacts on changing practice. Physicians will need to make trade‐offs in clinical practice to balance efficient resource utilization with the patient's evolving condition in the inpatient setting, the complexities of clinical workflow, and the patient's expectations.[14] Appropriate telemetry utilization is a prime example of what needs to be done well in the future for high‐value care.