The value of process friction: The role of justification in reducing medical costs

• Date: 01 January 2020
• Author: Anita L. Tucker,Jillian A. Berry Jaeker
• DOI: http://doi.org/10.1002/joom.1024
• Published date: 01 January 2020

RESEARCH ARTICLE

The value of process friction: The role of justification in reducing

medical costs

Jillian A. Berry Jaeker | Anita L. Tucker

Questrom School of Business, Boston

University, Boston, Massachusetts

Correspondence

Jillian A. Berry Jaeker, Questrom School of

Business, Boston University,

595 Commonwealth Avenue, Boston, MA

02215.

Email: jjaeker@bu.edu

Handling Editors: Lawrence Fredendall,

Anand Nair, Jeffery Smith

Abstract

We study the role of process friction in increasing efficiency of service provisions.

We examine one potential lever for reducing the provision of discretionary ser-

vices: “justification”—an otherwise non-value-added process step that introduces

process friction by forcing workers to explain the rationale for requesting an

optional service. We exploit the presence—and absence—of a justification step in

the ultrasound (US) ordering process at two emergency departments (EDs). We

find that patients with abdominal pain are less than half as likely to receive an US

when there is a justification step compared to when there is not. Additionally, we

find a spillover effect: other diagnostic tests are also ordered less frequently. The

decrease in testing reduces the average length of stay of the patients, and reduces

testing costs by more than $200,000, with no decrease in quality. We show that

two mechanisms underlie these results: (a) justification reduces clinicians' avail-

able time, and (b) justification forces clinicians to reflect on a patient's need for ser-

vice. Our paper contributes to recent theory on friction and reflection as drivers of

efficiency in services. We show that justification can serve as an effective lever for

reducing medical tests—and costs—without negatively impacting quality.

KEYWORDS

behavioral operations, health care, medical cost, performance improvement, worker discretion

1|INTRODUCTION

Credence-good providers—physicians, lawyers, consultants—

have substantial discretion over the services that customers

receive. Credence good customers risk being overserved with-

out a measurable increase in quality (Debo, Toktay, & Van

Wassenhove, 2008). We examine one such setting—

physician ordering of optional medical tests and services

(hereafter referred to as “tests”). Unnecessary tests escalate

expenses for patients and payers (Abaluck, Agha, Kabrhel,

Raja, & Venkatesh, 2016; Kharbanda et al., 2013; Schwartz,

Landon, Elshaug, Chernew, & McWilliams, 2014), and repre-

sent a substantial portion of health care costs in the United

States (Berwick & Hackbarth, 2012). The Institute of Medi-

cine (IOM) estimates that unnecessary services cost more than

$200 billion per year in the United States and states that these

extra costs “hinder progress in improving health and threaten

the nation's economic stability and global competitiveness”

(Medicine, 2013). There is increasing concern about such

inefficiencies in the U.S. health care system (Garber & Skin-

ner, 2008).

Information from a test can be more accurate than infor-

mation physicians can obtain themselves, such as by taking

a detailed patient history (American College of Emergency

Physicians, 2008). However, overuse may occur because,

for the ordering physician, these tests may provide non-

diagnostic benefits that inflate their use: higher fees, more

satisfied patients (Gawande, 2009a, 2009b), or a workload

reduction from shifting diagnostic work to another depart-

ment (Chan, 2015). Moreover, many physicians order tests

Received: 27 November 2017 Revised: 10 February 2019 Accepted: 2 March 2019

DOI: 10.1002/joom.1024

12 © 2019 Association for Supply Chain Management, Inc. wileyonlinelibrary.com/journal/joom J Oper Manag. 2020;66:12–34.

as a type of “defensive medicine”to reduce the probability

of facing a lawsuit (Kessler & McClellan, 1996; Studdert

et al., 2005). The concerns associated with unnecessary diag-

nostic tests include financial cost, potential for side effects,

patient discomfort, and false-positive findings that may lead

to unnecessary procedures (Griner & Glaser, 1982).

Decoupling physician payment from the tests ordered is

one lever to reduce test overuse (Gawande, 2009a, 2009b).

However, financial mechanisms are unlikely to solve the

problem completely. Even when physicians are not directly

compensated for the tests they order, the number of tests

ordered varies widely across physicians (Stiell et al., 1997).

Consequently, reducing test use will likely require non-

financial levers that motivate more prudent choices (Institute

of Medicine, 2001).

We investigate justification as one possible nonfinancial

lever for influencing physician providers (“providers”). We

define justification as a process step that requires providers

to explain their rationale when they request that their cus-

tomers receive optional services from a specialist (“special-

ist”). Justification may reduce testing via two mechanisms:

(a) justification introduces a process friction that reduces

providers' available time to order tests (“processing time”)

and (b) justification requires that the provider reflect on why

a test is needed (“reflection”). Although we anticipate that

justification will reduce testing, a priori, the magnitude of

the effect is unclear. Furthermore, in a setting with heteroge-

neous patients, providers, and multiple kinds of tests, it is

unclear which types of patients and tests will be most

affected. Finally, it is unknown whether justification will

benefit ED performance because it introduces an otherwise

non-value-added step—a process friction—to test ordering,

and may encourage suboptimal testing modalities.

To investigate these issues, we analyze the effect of a

change in the process to order an ultrasound (US) at one of

two affiliated EDs within 5 miles of each other, staffed by the

same set of physicians. This change required physicians to

justify their request for an US for some patients but not for

other, clinically similar patients. Our sample includes 10,669

patient visits in which the primary complaint was abdominal

pain, a complaint for which US is a common test. We find

that the justification step more than halves the average proba-

bility that a patient with abdominal pain will receive an US,

reducing it from 19 to 8%. We conduct additional analyses to

investigate and find support for the two mechanisms we

hypothesize are driving the test reductions—processing time

and reflection. Moreover, we find a spillover effect; the US

justification step also reduces orders for other diagnostic

services.

We show that justification improves system performance

by reducing test use and smoothing the flow of patients

through the ED. We find that adding friction to a process can

be beneficial. Our work contributes to the newly emerging

theories of the operational impact of process friction (Bavafa,

Hitt, & Terwiesch, 2017; Terwiesch, Asch, & Volpp, 2016)

and reflection (Di Stefano, Gino, Pisano, & Staats, 2015). We

also highlight a boundary to the speed-quality trade-off theory

(Anand, Paç, & Veeraraghavan, 2011; Hopp, Iravani, &

Yuen, 2007) by showing that justification can increase both

speed and quality. Finally, we provide a managerial strategy

to harness the benefits of justification, and quantify the finan-

cial benefits of this strategy.

2|LITERATURE REVIEW

We draw on the research of service operations scholars who

examine the interactions among work system design, worker

behavior, and system performance.

2.1 |The impact of available processing time

In service settings, workers often have discretion over the

type and number of services provided to customers. Service

operations research proposes that workers face a trade-off

between the conflicting goals of overall system speed and ser-

vice quality. The general model of this speed-quality trade-off

is that a provider can improve quality for a current customer

by providing more services, but this comes at the cost of

decreased speed, consequently increasing waiting times for

incoming customers (Anand et al., 2011; Debo et al., 2008;

Delasay, Ingolfsson, Kolfal, & Schultz, 2015; Hopp et al.,

2007; Wang, Debo, Scheller-Wolf, & Smith, 2010).

Studies have shown that a provider's workload influences

her decisions regarding the quantity and type of services to

provide. If providers have available time, they provide more

services to each customer to increase quality (Delasay et al.,

2015; Hopp et al., 2007). Conversely, if there is a high load

on the system, they speed up by providing less service to each

individual (George & Harrison, 2001; Stidham & Weber,

1989). Researchers have found that in practice, when work-

load is visibly high, workers “cut corners”(Oliva & Sterman,

2001) and work at a faster pace (KC & Terwiesch, 2012;

Schultz, Juran, Boudreau, McClain, & Thomas, 1998;

Shunko, Niederhoff, & Rosokha, 2017; Song, Tucker, &

Murrell, 2015; Song, Tucker, Murrell, & Vinson, 2015), and

that congestion-induced speed often comes at the expense of

quality (Anderson, Golden, Jank, & Wasil, 2012; KC & Ter-

wiesch, 2012; Kuntz, Mennicken, & Scholtes, 2014; Oliva &

Sterman, 2001).

While, in theory, the speed-quality trade-off can be opti-

mized (Chan, Farias, Bambos, & Escobar, 2012), in practice,

this proves difficult. During highly congested periods in hos-

pitals, physicians often cut corners by discharging patients

early, which results in a higher mortality rate (Kuntz et al.,

BERRY JAEKER AND TUCKER 13