Evaluating Time in Health Care: What Are We Busy About?
It's not enough to be busy. The question is: what are we busy about?" --Henry David Thoreau
Twenty years ago, I (Jodi Polaha) had a freshly minted PhD in Child Clinical Psychology and a strong desire to be a Mental Health Hero of Rural and Underserved Places. I was living in Nebraska, a place of great opportunity for a woman with such a wish. When a pediatric primary care practice offered me a room in their clinic free of charge, I was "all in." Before long, however, I realized that by colocating care, I was providing care that went above and beyond traditional mental health services.
My colleagues and I devised a rudimentary study to prove we were on to something special (Cooper, Valleley, Polaha, Begeny, & Evans, 2006). Our research assistants observed over 300 pediatric primary care clinic visits, collecting data on pretty much everything we thought might be interesting. It was the kind of "every variable" study that, now, as a journal editor, leaves me shaking my head.
We got lucky though. Within the mess of variables were the total number of concerns and behavioral health concerns raised, as well as the length of the visit. These data jumped out at us: parents or providers raised concerns about behavioral, developmental, or emotional well-being in 24% of all visits even though only 9% were identified as "psychological consultation" in advance. Visits in which provider or parent raised a psychosocial concern lasted 5-7 min longer than visits without such concerns, on average. We figured that in a busy practice where pediatricians might see 30-40 patients per day, these scenarios could translate to nearly an hour of unplanned extra time.
The time challenge in primary care is longstanding. A literature review shows physicians have reported struggles with having enough time to address patient concerns since one study conducted in 1966, if not earlier (Mechanic, 2001). Three decades later, deGruy (1997) published a seminal article in this journal pointing to behavioral health care as a key demand on time in primary care. Our 2006 article (Cooper et al., 2006) was one of the first to provide the evidence for that point; evidence the field clearly desires as, according to Google Scholar, it has been cited 124 times. Another study estimated that physicians would need 40 extra minutes per patient to address USPSTF preventative guidelines, many of which relate to behavioral health (Yarnall, Pollak, 0stbye, Krause, & Michener, 2003). Indeed, these data support another key premise in the deGruy article: that integrating behavioral services provides a better way to use everyone's time (deGruy, 1997).
It has been over 10 years since we published that article, and "time data" is still a keen interest in the field. Our editorial this month focuses on the evaluation of time in primary care including strategies for measurement, methods for estimating the value of "saved time," and a discussion about the challenges of measuring the indirect benefits of time-savings. Ultimately, a sophisticated evaluation of time and its value can provide a stronger business case supporting the uptake and sustainment of integrated care. Time is an important outcome of implementation, accessible to those in real world settings, and we hope to inspire more study in this area. Such work is consistent with our vision for Families, Systems, & Health (Polaha & Sunderji, 2018).
Measuring Time in Primary Care
In the 2006 study described above, student research assistants were assigned to follow one physician for 1-4 hr, using a stopwatch to record the start and end of each of 302 visits in total (Cooper et al., 2006). Research assistants were trained by coding a sample audio tape and three-four visits in real time, for which they reliably coded with at least 90% interrater reliability before independently coding. All of their observation sessions were audio-recorded, and 25% were randomly selected and coded independently. The length of the visit was coded as "in agreement" if the two measurements were within 1 min of each other. This method yielded a 97.4% agreement rate for this study.
Time-and-motion methods are common in health care and viewed as a gold-standard in clinical trials. One approach is to place observers outside exam rooms where they might observe the start and stop of multiple clinic visits at one time (Gouge, Polaha, Rogers, & Harden, 2016). This procedure can allow for the collection of time data on multiple providers at one time. It provides a great efficiency, if the physical design allows it. Other approaches include having research assistants follow the patient through the clinic visit, which can allow for an analysis of time spent on particular activities during the visit (e.g., Pizziferri et al., 2005), or collecting audio/video recordings, which can allow for fine-grained analysis of times associated with specific actions in a clinic visit (Tai-Seale, McGuire, & Zhang, 2007).
There are a number of advantages to time-and-motion methods. First, it is rather straightforward to operationally define the "start" and "stop" of visits (e.g., entering and final exit from room) for coding purposes. Studies that assess interrater reliability of these approaches have found that they produce consistent results (Cooper et al., 2006; Gouge, Polaha, & Powers, 2014). Another advantage is that volunteers may be easy to find: undergraduate research assistants love this kind of study, especially premed students who are thrilled to sit through health care visits. The downsides of this method are that it is time-intensive to collect data on a large number of participants, labor intensive (to find students, train them, create schedules for observation, and assess reliability, etc.), and can be intrusive for patients and providers. Another limitation is that these assessments of behavior can lead to reactivity among providers, a well-established influence on study outcomes. For example, if providers are aware that their behavior is being monitored, they could attempt to be more comprehensive in their procedures or attentive to patient concerns that usual thereby increasing total time of the encounter. Alternatively, if providers are aware that time is the focus of the observation, they may be inclined to move more quickly than usual through the appointment. It is, therefore, unknown to what extent this observer influence might impact time assessment in primary care.
Time Stamp in Electronic Health Record (EHR)
In March 2019, this journal published a brief report in which investigators conducted a retrospective chart review of all patient encounters over a 4-month period including electronic timestamps of the total visit duration as well as the portion devoted to medical care and the portion for behavioral care (Riley, Paternostro, Walker, & Wagner, 2019). In this way, they were able to demonstrate that while the total visit length (medical + behavioral) was longer when a behavioral health concern was addressed, the physician-only portion was significantly shorter when a behavioral health consultant was introduced into the visit (e.g., warm hand-off)- Authors described timestamps as "artifacts of clinical staff marking each phase of a patient visit...in the EHR in real-time..." (p. 164), deriving the method from a validated approach for workflow optimization described by Hribar and colleagues (2015). Their study demonstrates the potential for this method with a distinct advantage over using onsite observers: large-scale investigations are more feasible using informatics and more within reach for a wide range of clinics.
Are timestamps sensitive enough to measure the impact of behavioral health integration? Maybe. Another validation study by Hribar and colleagues (2018) showed that timestamps were within 3 min of observed times for more than 80% of clinic visits. That said, many time-stamps are only estimates of actual time spent. For example, my colleague Dr. Leigh Johnson and I just completed a study of medical assistants working as scribes in primary care. Specifically, scribes performed nursing duties at the beginning and end of the patient visit (e.g., vitals, ordering labs) but also stayed in the room for the duration of the visit, completing documentation while the resident physician focused on the patient's exam. A time stamp associated with the opening and closing of the patient's note showed "time to note closure" was 8.5 hr shorter when a visit was scribe-facilitated. These data suggest that residents spent less time on notes, but it is by inference only. Riley and colleagues (2019) suggest researchers prospectively develop a more rigorous approach by designing time stamps into EHR workflows in ways that allow for ease of extraction and analyses around key visit components (and less inference).
EHRs may not be the only source of technology we can leverage. Emerging technology-based approaches such as body cameras or sensors placed in exam rooms can track personnel entering and leaving. New technologies may offer promise in their efficiency as well as their potential to garner data from virtually any clinician behavior over the course of a workday to render a more complete assessment of how time is spent.
Other Measurement Methods
Time-and-motion studies and EHR time stamps require resources that may be unavailable outside of grant funding or other investments, or infeasible in some primary care-based settings. There are other rigorous approaches and methods available to measure time that practitioners and researchers can use, including structured surveys, patient or provider interviews, and administrative data sources. In the economic evaluation literature, one widely used method is a structured survey in which clinical leaders estimate and report time spent by each of various providers involved in a given intervention (Zarkin, Dunlap, & Homsi, 2004). This method allows for input regarding not only direct time spent but also indirect sources such as "outside" administrative time. A benefit of structured surveys is that they can usually be completed by a single person at a clinic relatively quickly. Using a survey can also facilitate data collection from multiple clinics in the context of larger health systems. However, surveys may not be as precise and tend to overestimate the time spent on service delivery due to recall bias and "heaping"--we are more likely to report time in minutes and at 5- or 10-min intervals. One workaround is that a structured survey can also be converted into a semistructured interview for clinical leaders and practitioners to increase precision.
Why Pick Just One?
Research is not a one-size-fits-all approach and sometimes we need to be creative in piecing together pieces of the puzzle by use time-and-motion and other approaches together--a "hybrid" approach. As an example, in one study of screening and brief intervention for illicit drug use in a primary clinic, a small sample of providers were trained to use a time diary to measure their time screening patients and conducting patient support activities (e.g., finding patients, EHR charting) and in addition, all brief interventions were recorded and time-stamped (Zarkin, Bray, Hinde, & Saitz, 2015). Having providers time themselves may not be as rigorous as having a trained observer collect time information but it balanced precision and feasibility within the available resources. Before collecting time data, providers estimated their patient support activities took 20-25 min, whereas the time diaries yield an average of 9.5 min.
Saving Time: For Whom and to What End?
In 2016 my colleagues and I published another "time paper" based in a rural pediatric practice. Using the time-and-motion method, we provided further evidence that providers spend more time in visits when the concerns raised include a psychosocial concern (as reported by the physician to the research assistant after each visit) (Gouge et al., 2016). In addition, we showed that when behavioral health consultants (BHCs) were available for warm hand-offs, providers spent about 2 min less per patient on average for every patient seen. The study went one step further: we showed that on days with behavioral health consultants available, the medical providers at the practice were able to see 42% more patients, collecting $1,142 more revenue than on days when no consultant was present. It is important to note that the BHCs in this study were advanced graduate students who were paid a stipend by the practice and did not bill for services (nonlicensed), so the increased revenue can be attributed to an increased capacity for medical provider billing (in theory due to having more time available to see more patients). In fact, Gouge, the lead investigator used these data to make a pitch for a different pediatric practice to hire her.
This example provides a fairly simple and compelling financial analysis, but falls short of a comprehensive evaluation of costs associated with the start-up and maintenance of an integrated care program, and of the clinical value of integrated care. Revenue generation and developing a self-sustaining financial model is critical to making the "business case" for clinical and administrative leadership but does not adequately reflect the true economic impact of integrating care. This simple case shows how one can substitute student labor to reduce medical provider time with patients. While the bottom line is important, there are broader outcomes and value judgments that need to be considered for integrated care.
A more comprehensive economic evaluation would account for training costs, for clinical supervision costs for BHCs, and for other administrative activities associated with adding BHCs to the clinic workflow, as well as the additional time spent by BHCs with patients and the reduced medical provider time. Integrating care inherently means adding new staff and processes to a clinic and spending more time with patients--this is more costly and results in higher costs per patient. Could this time be better spent on other activities? Economists often refer to this notion as "opportunity costs." Ultimately, we need to weigh whether the increased time and costs associated with integrated care result in better clinical or patient outcomes at an acceptable cost through a cost-effectiveness analysis. Cost-effectiveness analysis compares the costs and outcomes of two interventions and can be interpreted as the cost to increase an outcomes by one unit for one intervention relative to the other. Rather than focusing just on cost-savings, it is perhaps more important to demonstrate whether integrated care improves outcomes at relatively low cost.
Ultimately, time saved can translate into a wide range of valued practice and personal activities and outcomes. Key among those outcomes is the potential for reduced provider burn-out. It is probably intuitive that if providers save enough time to keep their lunch breaks, go home on time, avoid documentation in the evenings, or attend a yoga class, they might report greater satisfaction in their jobs. Apart from the benefit of these alternative ways of spending time, there is also a sense of well-being that comes from staying on schedule and feeling in control of one's time and clinic. To the extent that time savings may prevent burnout, there may also be an indirect cost savings associated with reduced staff turnover: it is expensive to search for and onboard new people. Time savings can result in other valued practice activities as well. In training settings such as residencies or internship programs, the savings can be translated into more time for instruction (Gouge et al., 2014). In our Family Medicine clinics we often talk about how instructors who have more time to engage students "sell" integrated care or Family Medicine to prospective providers. Finally, effective time distribution in team-based care means that everyone is working "at the top of their license," and there is a sense that providers are performing at their best.
At the end of the day, there are both economic and less tangible benefits to having predictable clinic operations in which people's medical and behavioral health needs are met. These different benefits, stemming from changes in how time is used, are relevant to a wide range of stakeholders including administrators, clinicians, and patients. In short, time is one of our most important resources in health care. Therefore, time studies have a crucial role to play in advancing the implementation of integrated care. In this editorial we describe several methods for measuring time and invite readers to consider which of these (or another method you're aware of) balances your needs for precision and feasibility of measurement.
Cooper, S., Valleley, R. J., Polaha, J., Begeny, J., & Evans, J. H. (2006). Running out of time: Physician management of behavioral health concerns in rural pediatric primary care. Pediatrics, 118, e132-e138. http://dx.doi.org/10.1542/peds.2005-2612
deGruy, F. V. (1997). Mental healthcare in the primary care setting: A paradigm problem. Families, Systems, & Health, 15, 3-26. http://dx.doi.org/10.1037/h0089802
Gouge, N., Polaha, J., & Powers, R. (2014). Bringing a behavioral health consultant to residency: Implications for practice and training. International Journal of Health Sciences Training, 2, 1-10.
Gouge, N., Polaha, J., Rogers, R., & Harden, A. (2016). Integrating behavioral health into pediatric primary care: Implications for provider time and cost. Southern Medical Journal, 109, 774-778. http://dx.doi.org/10.14423/SMJ.0000000000000564
Hribar, M. R., Read-Brown, S., Goldstein, I. H., Reznick, L. G., Lombardi, L., Parikh, M.,... Chiang, M. F. (2018). Secondary use of electronic health record data for clinical workflow analysis. Journal of the American Medical Informatics Association, 25, 40-46. http://dx.doi.org/10.1093/jamia/ocx098
Hribar, M. R., Read-Brown, S., Reznick, L., Lombardi, L., Parikh, M., Yackel, T. R., & Chiang, M. F. (2015). Secondary use of EHR timestamp data: Validation and application for workflow optimization. AMIA Annual Symposium Proceedings Archive, 2015, 1909-1917.
Mechanic, D. (2001). How should hamsters run? Some observations about sufficient patient time in primary care. British Medical Journal, 323, 266-268. http://dx.doi.org/10.1136/bmj.323.7307.266
Pizziferri, L., Kittler, A. F., Volk, L. A., Honour, M. M., Gupta, S., Wang, S.,...Bates, D. W. (2005). Primary care physician time utilization before and after implementation of an electronic health record: A time-motion study. Journal of Biomedical Informatics, 38, 176-188. http://dx.doi.org/10.1016/j.jbi.2004.11.009
Polaha, J., & Sunderji, N. (2018). A vision for the future of families, systems, & health: Focusing on science at the point of care delivery. Families, Systems, & Health, 36, 423-426.
Riley, A. R., Paternostro, J. K., Walker, B. L., & Wagner, D. V. (2019). The impact of behavioral health consultations on medical encounter duration in pediatric primary care: A retrospective match controlled study. Families, Systems, & Health, 37, 162-166. http://dx.doi.org/10.1037/fsh0000406
Tai-Seale, M., McGuire, T. G, & Zhang, W. (2007). Time allocation in primary care office visits. Health Services Research, 42, 1871-1894. http://dx.doi.org/10.1111/j.1475-6773.2006.00689.x
Yarnall, K. S. H., Pollak, K. I., Ostbye, T., Krause, K. M., & Michener, J. L. (2003). Primary care: Is there enough time for prevention? American Journal of Public Health, 93, 635-641. http://dx.doi.org/10.2105/AJPH.93.4.635
Zarkin, G, Bray, J., Hinde, J., & Saitz, R. (2015). Costs of screening and brief intervention for illicit drug use in primary care settings. Journal of Studies on Alcohol and Drugs, 76, 222-228. http://dx.doi.org/10.15288/jsad.2015.76.222
Zarkin, G. A., Dunlap, L. J., & Homsi, D. G. (2004). The substance abuse services cost analysis program (SASCAP): A new method for estimating drug treatment service costs. Evaluation and Program Planning, 27, 35-43. http://dx.doi.org/10.1016/j.evalprogplan.2003.09.002
Received October 7, 2019
Accepted October 7, 2019
Jodi Polaha, PhD
East Tennessee State University
Gregory P. Beehler, PhD, MA
VA Western New York Healthcare System, Buffalo, New York
Jesse M. Hinde, PhD
RT1 International, Research Triangle Park, Durham, North Carolina
Nadiya Sunderji, MD, MPH
Waypoint Centre for Mental Health Care, Penetanguishene, Ontario, Canada, and University of Toronto
Jodi Polaha, PhD, Department of Family Medicine, Quillen College of Medicine, East Tennessee State University; Jesse M. Hinde, PhD, Social Policy, Health and Economics Research Unit, RTI International, Research Triangle Park, Durham, North Carolina; Gregory P. Beehler, PhD, MA, VA Center for Integrated Healthcare, VA Western New York Healthcare System, Buffalo, New York; Nadiya Sunderji, MD, MPH, Waypoint Centre for Mental Health Care, Penetanguishene, Ontario, Canada, and Li Ka Shing Knowledge Institute and Department of Psychiatry, University of Toronto.
Correspondence concerning this article should be addressed to Jodi Polaha, PhD, Department of Family Medicine, Quillen College of Medicine, East Tennessee State University, P.O. Box 70621, Johnson City, TN 37614. E-mail: firstname.lastname@example.org
|Printer friendly Cite/link Email Feedback|
|Author:||Polaha, Jodi; Beehler, Gregory P.; Hinde, Jesse M.; Sunderji, Nadiya|
|Publication:||Families, Systems & Health|
|Date:||Dec 1, 2019|
|Previous Article:||Relationships Matter in Population Health.|
|Next Article:||Implementation Findings From an Effectiveness-Implementation Trial of Tablet-Based Parent Training in Pediatric Primary Care.|