Author	Affiliation
Adam L. Sharp, MD, MS	Southern California Kaiser Permanente, Department of Research and Evaluation, Pasadena, California
Tammy Chang, MD, MPH, MS	University of Michigan, Robert Wood Johnson Foundation Clinical Scholars Program, Ann Arbor, Michigan
Enesha Cobb, MD, MTS	University of Michigan, Robert Wood Johnson Foundation Clinical Scholars Program, Ann Arbor, Michigan
Weyinshet Gossa, MD, MPH	University of Michigan, Department of Family Medicine, Ann Arbor, Michigan
Zachary Rowe,	Friends of Parkside, Detroit, Michigan
Lauren Kohatsu,	University of Michigan, Ann Arbor, Michigan
Michele Heisler, MD, MPA	University of Michigan, Robert Wood Johnson Foundation Clinical Scholars Program, Ann Arbor, Michigan

Introduction
Methods
Results
Discussion
Limitations
Conclusion
Supplementary information

INTRODUCTION

The emergency department (ED) is increasingly the healthcare setting where patients seek acute unscheduled care,¹ with visits outpacing population growth, especially among disadvantaged populations.² Expanding insurance coverage with healthcare reform will likely further increase this trend.³ However, nearly all of the information regarding ED utilization is gathered from patients in the ED itself or from administrative data related to ED visits.⁴–⁸ We know patients perceive acute care differently than providers,⁹ but our lack of knowledge regarding patients’ real-time decision-making makes it difficult to best meet their needs. This is particularly salient when discussing acute care needs of low-income populations who use the ED at higher rates than others.⁵

Community-based participatory research (CBPR) has been suggested as a strategy to improve the health and well-being of communities and eliminate health disparities.¹⁰ CBPR represents a potentially useful strategy to understand populations known to be frequent ED users and collecting real-time patient data could help inform efforts to improve acute care. Since 80% of adults in the U.S. send text messages, this presents one promising option to gather information in the context of a patient’s usual life, prior to seeking acute care.¹¹–¹³

The objectives of this pilot study were to: 1) understand when participants decide if healthcare is needed, primarily for low-acuity conditions; 2) describe where patients choose to seek care for different conditions; 3) examine factors influencing decision-making; and 4) examine how preferences and perceptions influence decision-making.

METHODS

We used a community-based participatory research (CBPR) strategy to conduct a mixed methods evaluation of real-time patient decision-making for common acute conditions. We formed a steering committee comprised of University of Michigan researchers, a member of the “Villages of Parkside” (TVP), an urban, government-subsidized community, and a representative of Friends of Parkside (FOP), a non-profit community organization affiliated with TVP. The steering committee guided the research project from inception to completion and met on a regular basis to design and plan the study, recruit and enroll participants, discuss data collection and assess results. The University of Michigan researchers conducted data analysis, with frequent meetings with the full steering committee to discuss the presentation and interpretation of results. Key principles of CBPR emphasizing equal partnership between all members and building capacity within communities guided this study. ¹⁰,¹⁴ Our mixed methods approach was designed to use quantitative data to inform our focus group discussion guide and assist in analyzing our qualitative findings consistent with the “following a thread” strategy.¹⁵,¹⁶

Participants were recruited by FOP by texting TVP members through a previously established database of cell phone numbers, distribution of flyers and word-of-mouth referrals. Only adults affiliated with TVP who had a primary care provider and text-messaging capabilities were eligible for participation. Two recruitment meetings were held by researchers to explain the study, collect demographic information, and obtain written informed consent. Due to the exploratory, hypothesis-generating nature of this research, we did not perform power calculations, and financial constraints only allowed for 20 participants.

 

Table 2. Hypothetical low-acuity vignettes texted to participants. Responses are stratifed by those who would choose to seek emergency department care (ED), primary care evaluation (PCP) or no care. Proportions may exceed 100% due to rounding.

	ED	PCP	No care
10 common low-acuity medical scenarios
Your stomach has been hurting since last night. You threw up twice today.	4 (5%)	18 (23%)	57 (72%)
You’ve had a sore throat for 4 days and feel sick.	12 (15%)	34 (42%)	36 (44%)
You’ve felt sick and had a fever for two days.	22 (27%)	22 (27%)	38 (46%)
You’ve had a cough, runny nose and headache for 3 days.	8 (10%)	20 (26%)	49 (64%)
You’ve had a throbbing headache for 3 hours.	9 (10%)	5 (5%)	78 (85%)
You have a red itchy rash on your legs, it has been there for 4 days.	10 (14%)	38 (52%)	25 (34%)
You hurt your back picking up a child 2 days ago and it still hurts to move.	18 (23%)	23 (30%)	36 (47%)
You slipped walking up the stairs and injured your knee. It is swollen and painful to walk.	38 (49%)	15 (20%)	24 (31%)
You’ve had a runny nose for 5 days and now your right ear is hurting.	11 (15%)	49 (65%)	15 (20%)
You slipped in the bathroom, injured your back, it hurts to lie down and when you bend over or twist.	12 (17%)	44 (62%)	15 (21%)
All low-acuity scenarios	144 (18%)	268 (34%)	373 (48%)
Scenarios designed to prompt a PCP visit
You need a flu shot for your new job.	7 (9%)	65 (82%)	7 (9%)
You have had a mole on your leg for 10 years and are now concerned it needs to be evaluated.	7 (9%)	45 (56%)	29 (36%)
All PCP scenarios	14 (9%)	110 (69%)	36 (23%)
Scenarios designed to prompt an ED visit
All of a sudden you can’t move your right arm or leg and you can’t speak normally.	72 (94%)	2 (3%)	3 (4%)
You fell down the stairs your head is bleeding, you are confused and you can tell your leg is broken.	60 (98%)	1 (2%)	0 (0%)
All ED scenarios	132 (96%)	3 (2%)	3 (2%)

 

We created ten hypothetical acute care vignettes, patterned after the ten most common acute care complaints seen in U.S. primary care clinics.¹⁷ Each vignette was designed to be texted to participants (e.g., “You’ve had a sore throat for 4 days and feel sick”). To assess the proportion of participants who would pursue care and where they would seek it, participants were asked to respond “ER” if they would seek care in the ED, “MD” if they would seek recommendations from their primary care provider (PCP), or “Nothing” if they would not seek care. Additionally, each participant was instructed to give a brief explanation of why s/he chose that response. To control for possible automated responses to our texts, we sent four anchoring vignettes: two designed to prompt an ED visit (stroke and severe trauma), and two designed to prompt a PCP visit (vaccination and nevus evaluation). We sent each participant two text messages per day over six weeks. Questions were sent at different times of day and categorized as either regular PCP “on hours” (Monday–Friday 8:00 AM to 5:00 PM) or “off hours” (Monday–Friday 5:00 PM–10:00 PM, Saturdays and Sundays). Participants were paid $1 for responding ER, MD or nothing to each text and another $1 for texting a brief explanation. Table 2 displays all of the vignettes texted to participants.

We used a focus group to understand survey responses and better comprehend the decision-making process for acute conditions. All participants were invited and encouraged to participate in the focus group. The focus group was led by a moderator (AS) and facilitated by an assistant moderator (WG), audio recorded and transcribed.¹⁸,¹⁹ We designed an interview guide to help moderators explore: 1) how participants decide if healthcare is needed; 2) how participants decide where to seek healthcare; 3) factors influencing decision-making; and 4) how preferences and perceptions of the ED and PCP clinic influence decisions (Appendix).

The focus group was transcribed verbatim by a professional transcriptionist and reviewed for accuracy. We used inductive qualitative techniques informed by thematic analysis.¹⁸,¹⁹ Transcripts were reviewed line by line to identify prominent concepts and ideas to draft preliminary coding categories. These initial findings were reviewed, coding categories were created, and themes were added and clarified as a team. Four researchers (TC, EC, WG, AS) engaged in an inductive process of reading and manually coding the transcript. Codes were further clarified and a codebook with definitions was developed.

The full steering committee reviewed results in frequent meetings and discussions, using memos to identify emerging themes and describe relationships among coding categories. The final coding scheme and analysis of the findings were reviewed, and disagreements were discussed until consensus was reached. We organized the results using the coding scheme structure and illustrated the themes with representative quotations. To increase the validity of our qualitative data, we performed “member checking” where the overall results of the study were presented to participants. Each participant indicated that the results included and accurately represented their viewpoints.

RESULTS

Our sample (n=20) comprised primarily of African American women (85%) and Medicaid recipients (60%), the majority of whom had visited the ED in the last 12 months ^{(Table 1)}. We received text survey responses from 18/20 recruited participants, as two lost cell phone service prior to the beginning of our data collection. We received responses for 72% of our text message vignettes with a similar proportion of responses sent “on hours” and “off hours.” Most commonly participants chose “nothing” (48% of vignettes) indicating they would not have sought care for the hypothetical text scenarios, next most often they would have sought primary care (34%) and less frequently (18%) ED care ^{(Table 2)}. This did not differ when stratified by time of day; in fact, more participants responded they would have sought ED care during “on hours” compared to “off hours.”

Table 1. Study population, n=20.

Ages (years)
Range	19 – 62
Average	34.3
Females	17 (85%)
Race/ethnicity
Black	20 (100%)
Education
<High school	4 (20%)
High school equivalent (GED)	2 (10%)
High school graduate	1 (5%)
Some college	11 (55%)
College graduate	2 (10%)
Work status
Student	4 (20%)
Employed	7 (35%)
Unemployed	9 (45%)
Insurance
Medicaid	12 (60%)
Private	5 (25%)
No insurance	3 (15%)
Primary care physician visits/year
0	4 (20%)
1 to 10	12 (60%)
11 to 20	3 (15%)
>20	1 (5%)
Emergency department visits/year
0	8 (40%)
1 to 5	11 (55%)
6 to 10	1 (5%)

 

Twelve of our participants contributed in the focus group. Five general themes were identified through inductive analysis of the focus group and assessment of open-ended text responses ^{(Table 3)}. These themes helped us to understand how participants decide if care is needed and where to seek it when necessary. It became clear through our text responses and focus group discussion that participants were not confident when they should seek medical care.

Table 3. Focus group themes with representative quotes to understand how patients decide if care is needed and where to seek it.

Participants considered the severity and length of symptoms to be the primary factors driving the need to seek care, but did not mention the types of symptoms as a key reason to receive medical evaluation. For example, one participant stated:

“If you were sick, like you really, really can’t take it, then you go; other than that it’s home remedies.”

After deciding to seek care, participants were unclear where they should seek evaluation. As one participant noted: “… it’s hard to tell… a lot of people don’t know.” They preferred seeing a physician with whom they had an established relationship, but reported difficulty arranging PCP appointments.

 

Participants perceived the ED as overcrowded, expensive and a venue for over-testing, stating:

“When you go to emergency, you get chest x-ray; you get CAT scan, you get all of that and you still get no solution. And then you get this bill.”

 

Participants also felt ED providers did not offer the time and attention they desired, by expressing:

“You’re there like three, four, five hours and then you go in there, they check you out for like two minutes and then you have been there for five hours to get a two minute result and they send you home.”

Though our study was not designed to assess urgent care centers, participants continually brought up this option for acute care. Overall, urgent care centers, while not specifically defined by participants, were described as sites of care not affiliated with a hospital or a PCP clinic and were perceived positively. Participants felt urgent care centers were more accessible than their PCP and had more predictable wait times than the ED, which translated into more time spent with the provider.

DISCUSSION

Our pilot study helps to understand how patients assess the need for care and where they choose to seek it. Participants were unable to define clear indications to seek care at an ED, with the exception of severe trauma. We found participants were more likely to stay home or seek care from a PCP, even during “off-hours,” than to seek ED care. The majority of research and discussion surrounding ED utilization emphasizes who should not be there, but does not clarify who should. Our study identifies a lack of clarity on the part of patients regarding when to seek care and where to receive it. These findings are consistent with previous studies showing many patients seek ED services for conditions that could be treated in other settings.²⁰

Contrary to previous studies, our participants were not more likely to seek medical care based on the time of day²¹ and did not seek ED care more often during “off-hours.”²² This may be a result of our real-time data collection method, demonstrating that time of day played less of a role than other considerations when deciding where to seek care. Also, though our study was not designed to evaluate urgent care centers, it became apparent that our participants preferred this site of care to the ED and a PCP clinic for common low-acuity conditions.

Our pilot project also demonstrates text messaging as a feasible and acceptable alternative to paper surveys, especially appropriate in understanding real time decision-making. Text messaging during random times of the day and evening might more closely simulate subject responses to sudden unexpected events, such as an illness or injury. This method may therefore better reflect the decision whether and how to seek medical care under real-life circumstances, than a paper survey in a quiet environment at an expected time.

More research is needed to understand if our findings hold true in other populations, but based on our results, emergency medicine has an opportunity to identify and establish clear indications for patients to seek care in an ED. Once a consensus is reached, this information can be disseminated to communities and patients to help them understand when acute care is needed in an ED versus another setting. It has been suggested that solutions may be best found looking at system failures instead of patient factors.²³ Therefore, collaborating with other specialties, policy makers and administrators may be necessary to clarify the role of the ED in the broader health system. Additionally, the role of urgent care clinics requires further evaluation and strategic planning to optimize our ability to care for common acute conditions. Lastly, despite the trend of increasing ED use, even patients known as frequent users of the ED may have an unfavorable impression of emergency department services; this requires future inquiry and understanding.

LIMITATIONS

Similar to any small study, the findings from our pilot study may not generalize to other communities and populations. Our findings need to be confirmed in different populations with larger samples. We used hypothetical vignettes to understand participant decision-making and it is possible that participants’ actual behavior could differ from their responses. Our study also used very brief hypothetical clinical vignettes, which may make it difficult to determine the seriousness of symptoms. Similarly, we did not send text messages between 10pm and 8am because our community partners felt this would not be well received by participants and text messages would likely not receive a response during these hours. Therefore, our study does not address behaviors regarding acute care visits during these times.

CONCLUSION

In our sample, participants were unclear when or where to seek care for common acute health problems, suggesting a need for future research and patient education regarding this matter. Our participants preferred the urgent care over going to the ED or PCP, and there is a need for a greater understanding of the role of the urgent care in acute care delivery.

Supplementary Information

Appendix PDF

Footnotes

Full text available through open access at http://escholarship.org/uc/uciem_westjem

Address for Correspondence: Allen Sharp, MD. MS, Department of Research and Evaluation, 100 S Los Robles, 2^nd Floor, Pasadena, CA 91101. Email: adam.l.sharp@kp.org. 9 / 2014; 15:675 – 681

Submission history: Revision received December 5, 2014; Submitted May 15, 2014; Accepted May 27, 2014

Conflicts of Interest: By the WestJEM article submission agreement, all authors are required to disclose all affiliations, funding sources and financial or management relationships that could be perceived as potential sources of bias. The authors disclosed none.

REFERENCES

1 Pitts SR, Carrier ER, Rich EC Where Americans get acute care: increasingly, it’s not at their doctor’s office. Health Affair. 2010; 29:1620-1629

2 Tang N, Stein J, Hsia RY Trends and characteristics of US emergency department visits, 1997–2007. JAMA. 2010; 304:664-670

3 Smulowitz PB, Lipton R, Wharam JF Emergency department utilization after the implementation of Massachusetts health reform. Ann Emerg Med. 2011; 58:225-234 e221

4 Northington WE, Brice JH, Zou B Use of an emergency department by nonurgent patients. Am J Emerg Med. 2005; 23:131-137

5 Cunningham PJ What accounts for differences in the use of hospital emergency departments across US communities?. Health Affair. 2006; 25:W324-W336

6 Ragin DF, Hwang U, Cydulka RK Reasons for using the emergency department: Results of the EMPATH Study. Acad Emerg Med. 2005; 12:1158-1166

7 Guttman N, Zimmerman DR, Nelson MS The many faces of access: reasons for medically nonurgent emergency department visits. J Health Polit Policy Law. 2003; 28:1089-1120

8 Young GP, Wagner MB, Kellermann AL Ambulatory visits to hospital emergency departments – Patterns and reasons for use. JAMA. 1996; 276:460-465

9 Langdorf MI, Bearie BJ, Kazzi AA Patients’ vs. Physicians’ Assessments of Emergencies: The Prudent Layperson Standard. Cal J Emerg Med. 2003; 4:75-81

10 Israel BA . Methods in community-based participatory research for health. 2005;

11 . Cell Phone Activities 2012. 2012;

12 Kew S Text messaging: an innovative method of data collection in medical research. BMC Res Notes. 2010; 3:342

13 Anhoj J, Moldrup C Feasibility of collecting diary data from asthma patients through mobile phones and SMS (short message service): response rate analysis and focus group evaluation from a pilot study. J Med Internet Res. 2004; 6:e42

14 Israel BA, Parker EA, Rowe Z Community-based participatory research: lessons learned from the Centers for Children’s Environmental Health and Disease Prevention Research. Environ Health Perspect. 2005; 113:1463-1471

15 O’Cathain A, Murphy E, Nicholl J Three techniques for integrating data in mixed methods studies. Bmj. 2010; 341:c4587

16 Creswell JW . Research design: qualitative, quantitative, and mixed methods approaches. 2009;

17 . National Hospital Ambulatory Medical Care Survey: 2009 Outpatient Department Summary Tables. 2009;

18 Kitzinger J Qualitative research. Introducing focus groups. Bmj. 1995; 311:299-302

19 Powell RA, Single HM Focus groups. Int J Qual Health Care. 1996; 8:499-504

20 Oster A, Bindman AB Emergency department visits for ambulatory care sensitive conditions: insights into preventable hospitalizations. Medical care. 2003; 41:198-207

21 DeLia D, Cantor JC, Brownlee S Patient preference for emergency care: can and should it be changed?. Med Care Res Rev. 2012; 69:277-293

22 Ginde RMCR, Kirzinger WK . Emergency room use among adults aged 18–64: Early release of estimates from the National Health Interview Survey. 2012;

23 Lowe RA, Localio AR, Schwarz DF Association between primary care practice characteristics and emergency department use in a medicaid managed care organization. Medical care. 2005; 43:792-800