Currently Viewing:
The American Journal of Managed Care Special Issue: Health Information Technology - Guest Editor: Farzad Mostashari, MD, ScM
Health Information Technology: On the Cusp of Healthcare Transformation
Ashish K. Jha, MD, MPH
The Data Revolution Comes to Healthcare
Farzad Mostashari, MD, ScM, Visiting Fellow, Brookings Institute, Former National Coordinator for Health IT, US Department of Health and Human Services
The Impact of Electronic Health Record Use on Physician Productivity
Julia Adler-Milstein, PhD; and Robert S. Huckman, PhD
Currently Reading
Face Time Versus Test Ordering: Is There a Trade-off?
James E. Stahl, MD, CM, MPH; Mark A. Drew, BID; Jeffrey Weilburg, MD; Chris Sistrom, MD, MPH, PhD; and Alexa B. Kimball, MD, MPH
Financial Effects of Health Information Technology: A Systematic Review
Alexander F. H. Low, MBA; Andrew B. Phillips, RN, PhD; Jessica S. Ancker, MPH, PhD; Ashwin R. Patel, MD, PhD; Lisa M. Kern, MD, MPH; and Rainu Kaushal, MD, MPH
Employing Health Information Technology in the Real World to Transform Delivery
Marsha Gold, ScD
Redesigning the Work of Case Management: Testing a Predictive Model for Readmission
Penny Gilbert, MBA, BSM, BSN, RN, CPHQ; Michael D. Rutland, MBA, FHFMA, FACHE, FABC; and Dorothy Brockopp, PhD, RN
Small Practices' Experience With EHR, Quality Measurement, and Incentives
Rohima Begum, MPH; Mandy Smith Ryan, PhD; Chloe H. Winther, BA; Jason J. Wang, PhD; Naomi S. Bardach, MD; Amanda H. Parsons, MD; Sarah C. Shih, MPH; and R. Adams Dudley, MD, MBA
Older Adult Consumers' Attitudes and Preferences on Electronic Patient-Physician Messaging
Richard Lam, MD, MBA; Victor S. Lin BS; Wendy S. Senelick, MPH; Hong-Phuc Tran, MD; Alison A. Moore, MD, MPH; and Brandon Koretz, MD, MBA
Will Meaningful Use Electronic Medical Records Reduce Hospital Costs?
William E. Encinosa, PhD; and Jaeyong Bae, MA

Face Time Versus Test Ordering: Is There a Trade-off?

James E. Stahl, MD, CM, MPH; Mark A. Drew, BID; Jeffrey Weilburg, MD; Chris Sistrom, MD, MPH, PhD; and Alexa B. Kimball, MD, MPH
Real-time location systems can capture face time and trade-offs between face time and diagnostic testing so that clinicians' responses to time pressures can be measured.
Background: As clinician-patient face time comes under pressure, clinicians might consider substituting testing for time spent in diagnostic reasoning, history, and physical exam.

Objectives: To explore the relationship between clinician-patient time and medical resource utilization.

Methods: In the Massachusetts General Hospital/Massachusetts General Physician Organization outpatient radio frequency identification project, clinicians and patients wore real-time location system (RTLS) tags. “Face time” was defined as the duration patients and clinicians were colocated. Radiology testing was used as a proxy for medical resource use. A radiology test was determined to be associated with a clinical encounter if it involved the same patient and clinician and occurred less than 3 months after the index encounter. Radiologic data were derived from the electronic health record and test appropriateness proxy score from the radiology order entry system. Data were synthesized and analyzed using standard structured query language and statistical analysis.

Results: From July 2008 to October 2010, 2086 clinical encounter medical records and RTLSmeasured face times could be associated: 1957 for primary care (PC) and 129 for urgent care (UC). Of these, 471 met study criteria. In PC, shorter face time was associated with more testing, but shorter wait times and flow times. In UC, testing was not associated with shorter face times, but was associated with shorter wait times and longer flow times.

Conclusion: Our pilot suggests RTLS can capture face time and trade-offs between face time and testing. Ongoing studies will elucidate how these trade-offs affect patients, clinicians, and healthcare systems.

Am J Manag Care. 2013;19(11 Spec No. 10):SP362-SP368
We can now successfully install real-time location systems (RTLSs), run these systems in healthcare settings, and extract useful information from them. We can capture work flow behavior such as face time, wait time, flow time, and utilization, and link these data to other sources such as electronic medical records for deeper insight into how our healthcare delivery systems behave. In this study, we demonstrated that:

  • We can measure face time, wait time, and flow time with RTLSs.

  • We can link clinical behavior (eg, face time) with resource utilization.

  • We can distinguish different clinical behavior across clinical units.
The time patients and clinicians spend together is usually when the most important clinical decision making occurs, is central to the clinician-patient relationship, and is at the core of clinical care.1-16 Time, unfortunately, is not an infinite resource. In clinical scheduling, time is often allocated using schedule templates with blocks of time, such as 15/30 minutes or 20/40 minutes. In these  limited intervals, the physician and patient must address a wide variety of concerns and complaints. This set of complaints and concerns (ie, the chief complaint, secondary, tertiary, and others) take time to address. Recent work suggests that within an encounter a clinician typically deals with one of these problems every 1 to 3 minutes.17 In this way, the clinical encounter might be considered a packing problem. A packing problem is a class of optimization problems where the object is to pack objects together  as densely as possible18 (also often referred to as knapsack or pallet problems). In the case of the clinical encounter, a packing problem refers to how much useful activity (eg, addressing clinical complaints) a clinician and patient can “pack” into their limited time together.

If time is limited, how do clinicians respond? As encounter time shrinks, as it may do under volume-based incentive structures, the number of complaints that can be managed within the allocated time interval also decreases. However, clinicians do not disregard or drop complaints randomly. Rather, dealing with the chief complaint is usually preserved at the potential expense of postponing  addressing other concerns perceived as less pressing.17,19,20 This strategy may be inadequate in the long run, in that problems deferred can ripen and return in a less manageable, more costly state.

There are a range of strategies one might pursue when dealing with diagnostic challenges as encounter time decreases. A common strategy used in many industries that deal with overtaxed resources is to try to use these resources more efficiently or off-load some of the demand to parallel resources. In the clinical setting, a way to off-load the demands of decision making and diagnostic reasoning may be to be to replace these time-intensive processes with testing or referrals, which may be done offline or asynchronously.

Measuring the face time, the part of the clinical encounter where the clinician and patient are actually together, is challenging in the clinical setting. Previous methods have ranged from surveys and administrative data to time-motion studies.

Each method has had its own set of limitations.7,8,21-26 Surveys and administrative data are subject to recall bias and anchoring or averaging of values. The latter is a particular problem because it is clearly known that variance is at least as important as average behavior, if not more so, in driving system performance.27 Timemotion studies are excellent at identifying specific measures in detail but are expensive and prone to distort subject behavior just by the presence of observers.

Real-time location systems (RTLSs) are a relatively new method that can help solve this problem. Commonly used outside of healthcare, RTLSs are a class of technologies used to identify and track the location of objects, typically in an indoor setting. Metaphorically they may be considered an indoor form of a global positioning system. As such, they hold similar potential to be used in a wide variety of ways and to serve as the infrastructure for yet-to-be-developed innovations beyond inventory and people tracking. These systems can be based on a wide variety of signaling modalities including radiofrequency (RF), infrared (IR), and ultrasound,28 and use a variety of methods to resolve the location of an object or person in time and space. In active systems, the tags themselves emit a signal that is detected by an array of receivers. In passive systems, the antennae/receivers emit a signal in a mode such as RF, the tags reflect it, and the receiver detects it. All systems have varying degrees of granularity, signal resolution, noise, and latency with regard to location and time. To date, the most common use of RTLSs both in and out of healthcare has been for inventory tracking.29-31 Part of the purpose of the Massachusetts General Hospital/Massachusetts General Physician Organization (MGH/MGPO) outpatient radio frequency identification (RFID) project has been to determine the feasibility of adapting these systems for use in busy clinical settings and monitor clinic system behavior.32,33

In this analysis we were able to use RFID to determine the trade-off between face time and diagnostic testing.


As part of the MGH/MGPO outpatient RFID project, several primary care, urgent care, and specialty clinics had RFID-based RTLSs installed. Clinicians and patients were asked to wear RTLS tags. Participation was voluntary. Flow data were derived from RTLS. Wait time was defined as the time from check-in/registration to the time the patient and clinician were colocated in the same exam room. Face time as defined as the length of time that patients and clinicians were colocated in the same room. Flow time was defined as the time from patient check-in/registration to patient checkout/ deregistration. See Figure 1. This project was approved by the MGH institutional review board. 

Radiology testing was used as a proxy for medical resource use. These tests are part of routine medical practice and are relatively easy to identify from the electronic health record (EHR). The inclusion criterion was any radiology test for an adult patient that could be associated with a specific clinical encounter. This condition was met if the radiology test shared the patient’s unique identifier, if the ordering provider and face-time clinician were the same, and if it was the first radiology test to occur less than 3 months after the index encounter. Patient under 21 years old were excluded from the study. Test data were derived from the hospital research patient data registry and EHR.

Testing Appropriateness

The appropriateness of the tests was estimated using the MGH radiology order entry (ROE) system.34 Radiology tests ordered via ROE receive a score that is a proxy for level of appropriateness (range: 1-9, where 1 is completely unacceptable and 9 is completely acceptable). The score is generated online at the time of order entry and has to be acknowledged by the ordering clinician before the order goes through. The scores in our system end to be skewed toward the high end, with a median value in our sample of 8. Tests with scores of 7 or more met the minimal level of appropriateness for this study. Clinicians can order radiology tests regardless of the score they receive.

Data were processed and stored in an encrypted secured database. Data were analyzed using JMP software (SAS Institute Inc, Cary, North Carolina) with standard statistical analytic methods.


From July 2008 to October 2010, 2086 clinical encounter medical records and measured face times could be associated. There were 1957 primary care visits and 129 urgent care visits (Figure 2). The median age of the patients was 45 years (range, 19-96 years); 35% were male and 65% were female.

Most (77%) patients were Caucasian, 7.8% were black, 7.3% were Hispanic, and 5.5% were Asian.


Out of 2086 encounters, 792 patients had associated radiology tests, 471 met study criteria, 368 had tests ordered through ROE, and 72 received ROE scores of 1 or higher. Scores ranged from 1 to 9. In primary care, 9% of visits resulted in radiology testing, whereas in urgent care 55% of encounters resulted in some form of radiology testing. Testing also occurred more frequently when encounters occurred in the fi rst half of the calendar year (P <.01) and when tests were ordered in second half of the workday (P = .1).

Process Measures

Testing Appropriateness. A total of 72 encounters could be associated with the EHR, the RTLS database, and the ROE database, and had scores of 1 or greater. See the Table for time versus testing results. On the whole, face time was weakly correlated with ROE score. However, when stratifi ed by the subgroups, primary care and urgent care providers behaved differently (Figure 3). As time increased for patients seeing urgent care providers, appropriateness increased. The opposite was true for the primary care group. In the primary care group, patients with the low-scoring tests (score <7) had both longer face time duration and more face time variance than patients  with higher-scoring tests. Patients with low-score tests had a mean face time duration of 6.5 minutes (standard deviation [SD], 9.2 minutes) versus a mean face time duration of 5.7 minutes (SD, 5.6 minutes) for patients with high-scoring tests (P = .01 for difference in variance). Of the low-scoring (<7) tests in both groups (9 in primary care, 5 in urgent care), there were 7 magnetic resonance imaging (MRI) studies of the spine (5 lumbar, 2 thoracic), 1 MRI study of the brain, 1 MRI study of the knee, 1 magnetic resonance cholangiopancreatography, 2 computed tomography (CT) scans of the brain, 1 CT scan of the abdomen, and 1 x-ray for a foreign body in the eye.


During the course of the MGH/MGPO RFID outpatient project, there were several periods in the development and implementation process when medical record numbers were not collected due to clinic preference. This incomplete data capture (<100%) could have led to sampling bias. That substantial data were gathered across 5 separate clinics over several time periods should help allay this concern.


The idea of real-time location has been a recurring theme in popular culture for a long time. It has regularly appeared in science fiction and has long been a technology on the verge of realization. Real-time location systems are technologies designed to let the end user answer the questions of what, where, and when: what it is you are looking for; where it is; and importantly, when it was last at that location (as close to real time as possible). That is in contrast to conventional inventory systems, which focus more on what and where with relatively little focus on temporal resolution, and on time-motion analysis methods, which tend to be very labor intensive and to have significant sampling issues.

Real-time location systems use a wide variety of physical modalities: RF (eg, Wi-Fi, ultra-wideband), IR, and ultrasound. These categories break down further into the functional modalities of passive, active, or hybrid, each with their own strengths and weaknesses.

Copyright AJMC 2006-2019 Clinical Care Targeted Communications Group, LLC. All Rights Reserved.
Welcome the the new and improved, the premier managed market network. Tell us about yourself so that we can serve you better.
Sign Up