540987
research-article2014
CPJXXX10.1177/0009922814540987Clinical PediatricsSturm et al
Article
Reconnecting Patients With Their Primary Care Provider: An Intervention for Reducing Nonurgent Pediatric Emergency Department Visits
Clinical Pediatrics 2014, Vol. 53(10) 988–994 © The Author(s) 2014 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/0009922814540987 cpj.sagepub.com
Jesse J. Sturm, MD, MPH1,2, Daniel Hirsh, MD3,4, Brad Weselman, MD5, and Harold K. Simon, MD, MBA3,4
Abstract Objective. Intervention to reduce nonurgent pediatric emergency department (PED) visits over a 12-month followup. Methods. Prospective, randomized, controlled trial enrolled children seen in the PED for nonurgent concerns. Intervention subjects received a structured session/handout specific to their primary care provider (PCP), which outlined ways to obtain medical advice. Visitation to the PED and PCP were followed over 12 months. Results. A total of 164 patients were assigned to the intervention and 168 patients to the control. At 12-month follow-up, the intervention group had a lower rate of nonurgent PED utilization compared with the control group (70 [43%] patients in the intervention compared with 91 [54%] in the control; P = .047). At 12 months, there was an increase in the rate of sick visits to PCP in the intervention group when compared with the control (P = .036). Conclusions. Intervention designed in cooperation with pediatricians was able to decrease nonurgent PED utilization and redirect patients to their PCP for future sick visits over a 12-month period. Keywords health care utilization, emergency department use
Introduction Each year, an estimated 20 million children in the United States seek medical care in pediatric emergency departments (PEDs). Utilization of the PED as a source of primary care is well documented in the literature.1,2 Multiple published studies indicate that up to half of these visits are for nonurgent complaints3-5 although a consensus definition of nonurgent has remained elusive.6 The use of the PED for nonurgent medical problems is considered both fiscally imprudent because of the high charges of such visits and medically undesirable because of the discontinuous and uncoordinated care received.7 Nonurgent visits that could be managed by the primary care provider (PCP) are also missed opportunities to provide essential childhood preventative care.8 Continuity of care,9-11 access to primary care/preventative care,12-14 and increased parental education about handling childhood illnesses15 all have been associated with decreased PED utilization and/or hospitalization. A literature review by Padgett and Brodsky16 and other studies reveal numerous factors associated with
nonurgent PED utilization. The most frequently cited factors influencing the nonurgent use of the PED are “the doctors office is closed,” followed by “my child was so sick and they required immediate attention,” “sent here by my doctor,” “the PED is closer to my home than the PCP office,” “child is seen more quickly in the PED,” and “I don’t know how to reach the PCP office when it is closed.”8,16,17 Prior to coming to the PED, patients may have the opportunity to contact nurse triage lines or directly contact their PCP to discuss their concerns.18 In cases where 1
Connecticut Children’s Medical Center, West Hartford, CT, USA University of Connecticut School of Medicine, Farmington, CT, USA 3 Children’s Healthcare of Atlanta, Atlanta, GA, USA 4 Emory University School of Medicine, Atlanta, GA, USA 5 Kids Health First Primary Care Network, Atlanta, GA, USA 2
Corresponding Author: Jesse J. Sturm, Connecticut Children’s Medical Center, 12 Spruce Lane, West Hartford, CT 06107, USA. Email [email protected]
Sturm et al after-hours referral to the PED is indicated by a nurse triage line, referring calls to a physician for a secondlevel triage reduced after-hours referral rates by 77%.19 This reduction was attributed to physician’s determination that the problems did not require urgent referral, prior knowledge about the family’s ability to care for the patient, knowledge of the patients past medical history, and change in the patient’s condition.19 Interventional studies have been published with the goal of decreasing nonurgent PED usage. Grossman et al20 studied nonurgent Medicaid patients in the PED and utilized nurses and social workers to work with families for 3 months to assign them a PCP, eliminate barriers to care (such as transportation), and ensure follow-up. This intensive intervention resulted in 14% fewer nonurgent PED visits in the subsequent 6 months, but no effect at 12 months.20 Similar studies using telephone/written reminders to follow-up with the PCP following PED visits and reminders sent to pagers/email show short-term improved follow-up and decreased PED utilization in intervention groups, which is not sustained after 3 to 6 months.21-24 To our knowledge, no published studies have employed a collaborative PCP-specific strategy, which emphasizes follow-up appointments and instructs patients on the proper steps that should be taken in the future for similar nonurgent concerns. Importantly, our intervention is constructed with direct collaborative input from the PCP and is practice specific. We hypothesize that this PCP-specific intervention, with a relatively low cost and minimally labor intensive design, has the potential to have sustainable impact on future nonurgent PED usage.
Methods Study Design and Location This is a prospective randomized intervention trial conducted at the PED of a tertiary care children’s hospital. Study enrollment took place from February 2010 to May 2010 from 8 am to 8 pm, 5 days a week. Eligible children were 3 months to 16 years old whose caregiver selfidentified a PCP that was within a network of 35 cooperative practices. This practice group is composed of independent physician groups who have a central office for data sharing. Each practice is independently operated and has differing levels of capabilities to obtain labs/radiographs, variable hours, and different afterhours availability. At the time of the study, there were 35 physician offices in the group, comprising 185 pediatricians. Previous published studies have shown that these practices account for approximately 60% of the nonurgent visitations to the study PED.25
989 Further eligibility criteria for enrollment were the following: presentation with a primary caregiver responsible for medical decision making and nurse triage acuity as low acuity (defined as one of the lowest 2 triage categories on the 5-level Emergency Services Index triage system). The 2 lowest triage levels, indicate a likelihood for low resource utilization.16,26-28 Previous studies have used a validated method to determine visit urgency, which defined visits that resulted in a radiograph, laboratory study, electrocardiogram, electroencephalogram, or admission as urgent visits; all other visits were defined as nonurgent visits.26,29-31 Our study will define nonurgent patients as those with a low initial triage acuity who did not have a radiograph, laboratory study (other than rapid strep), electrocardiogram, electroencephalogram, and were not admitted. If initially triaged as a nonurgent visit based on triage acuity, but subsequently tests were ordered by the treating physician, the patient was no longer eligible to be in the study and was removed. Eligible patients were approached and consented for enrollment in the study prior to physician encounter. To ensure standardized consent and intervention sessions, study personnel received a 60-minute training session and followed a scripted description of the study. All consent forms and study methodology were approved by the institutional review board. All enrolled patients completed a questionnaire in the waiting room about what resources they used prior to coming to the PED for care and identified barriers to obtaining timely care. This questionnaire also assessed patient baseline satisfaction with access to their PCP office (Supplemental Information 1). The surveys were developed in accordance with previous guidelines32 and included a literature review of existing research, expert panel item generation, reduction, and pretesting. The survey used dichotomous and frequency-of-endorsement formats. Based on even or odd medical record numbers, a random sample of enrolled patients was selected to participate in the intervention arm. The intervention took place at the conclusion of the visit, after being seen by the physician and prior to discharge from the PED. Participants selected for the intervention were given a laminated one-page handout developed with input from their self-identified PCP which outlined office hours/location, scope of practice (ability to do radiographs, blood work, sutures, etc), and the preferred steps that patients should take to address medical concerns (Supplemental Information 2). Intervention subjects were told that the form was developed with direct input from their PCP. In a standardized 10-minute session, study representatives described the algorithm to caregivers and answered questions. Intervention patients were given a laminated copy of this one-page
990 form to take home. Control patients were managed with the standard of care; they were given routine discharge instructions by the PED nurse. Intervention patients similarly received standard discharge instructions. Spanish-speaking patients had a translated copy of the questionnaire and educational intervention form as required. If needed, an on call translator was used to answer questions in Spanish. The PCP-specific form used for intervention was developed in coordination with each of the participating 35 practices. A sample form was completed by the lead physician at each practice. If a practice used a telephone call center for after hours advice, the forms noted that a physician was on call every night and if the patient wished they could ask the call center to contact the afterhours physician. Each practice outlined what they felt were the optimal strategies to obtain this medical advice and how to best obtain same day sick visits. For the intervention and control groups, future visitation to either of the 2 local PEDs for both acute and nonacute visits was tracked via the electronic medical record at 6- and 12-month intervals. These are the only dedicated tertiary care PEDs in the area, so any returns to PEDs would be expected to occur at one of these sites. Since enrollment occurred over a period of several months, patients reached the 6- and 12-month intervals at slightly different times; this was accounted for in data extraction. In addition to PED visitation, future visits to the PCP office for both sick visits and well-child visits were quantified at the end of the 12-month study by examining billing data from the PCP offices. At 6 months after study enrollment was complete, patients in the intervention group were contacted by telephone to assess their satisfaction with access to their PCP office. Contact was attempted on 3 separate occasions; if the enrolling caretaker could not be contacted, the questionnaire was mailed in an addressed envelope. The questionnaire was identical to the one completed at study entry. Results were scored on a Likert-type scale and comparisons were made across the groups over the 6-month period. To be consistent with previous research, the timeliness composites, scored on a 4-point Likerttype scale, were dichotomized, with scores of >3.5 indicating the highest quality care and scores of
research-article2014
CPJXXX10.1177/0009922814540987Clinical PediatricsSturm et al
Article
Reconnecting Patients With Their Primary Care Provider: An Intervention for Reducing Nonurgent Pediatric Emergency Department Visits
Clinical Pediatrics 2014, Vol. 53(10) 988–994 © The Author(s) 2014 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/0009922814540987 cpj.sagepub.com
Jesse J. Sturm, MD, MPH1,2, Daniel Hirsh, MD3,4, Brad Weselman, MD5, and Harold K. Simon, MD, MBA3,4
Abstract Objective. Intervention to reduce nonurgent pediatric emergency department (PED) visits over a 12-month followup. Methods. Prospective, randomized, controlled trial enrolled children seen in the PED for nonurgent concerns. Intervention subjects received a structured session/handout specific to their primary care provider (PCP), which outlined ways to obtain medical advice. Visitation to the PED and PCP were followed over 12 months. Results. A total of 164 patients were assigned to the intervention and 168 patients to the control. At 12-month follow-up, the intervention group had a lower rate of nonurgent PED utilization compared with the control group (70 [43%] patients in the intervention compared with 91 [54%] in the control; P = .047). At 12 months, there was an increase in the rate of sick visits to PCP in the intervention group when compared with the control (P = .036). Conclusions. Intervention designed in cooperation with pediatricians was able to decrease nonurgent PED utilization and redirect patients to their PCP for future sick visits over a 12-month period. Keywords health care utilization, emergency department use
Introduction Each year, an estimated 20 million children in the United States seek medical care in pediatric emergency departments (PEDs). Utilization of the PED as a source of primary care is well documented in the literature.1,2 Multiple published studies indicate that up to half of these visits are for nonurgent complaints3-5 although a consensus definition of nonurgent has remained elusive.6 The use of the PED for nonurgent medical problems is considered both fiscally imprudent because of the high charges of such visits and medically undesirable because of the discontinuous and uncoordinated care received.7 Nonurgent visits that could be managed by the primary care provider (PCP) are also missed opportunities to provide essential childhood preventative care.8 Continuity of care,9-11 access to primary care/preventative care,12-14 and increased parental education about handling childhood illnesses15 all have been associated with decreased PED utilization and/or hospitalization. A literature review by Padgett and Brodsky16 and other studies reveal numerous factors associated with
nonurgent PED utilization. The most frequently cited factors influencing the nonurgent use of the PED are “the doctors office is closed,” followed by “my child was so sick and they required immediate attention,” “sent here by my doctor,” “the PED is closer to my home than the PCP office,” “child is seen more quickly in the PED,” and “I don’t know how to reach the PCP office when it is closed.”8,16,17 Prior to coming to the PED, patients may have the opportunity to contact nurse triage lines or directly contact their PCP to discuss their concerns.18 In cases where 1
Connecticut Children’s Medical Center, West Hartford, CT, USA University of Connecticut School of Medicine, Farmington, CT, USA 3 Children’s Healthcare of Atlanta, Atlanta, GA, USA 4 Emory University School of Medicine, Atlanta, GA, USA 5 Kids Health First Primary Care Network, Atlanta, GA, USA 2
Corresponding Author: Jesse J. Sturm, Connecticut Children’s Medical Center, 12 Spruce Lane, West Hartford, CT 06107, USA. Email [email protected]
Sturm et al after-hours referral to the PED is indicated by a nurse triage line, referring calls to a physician for a secondlevel triage reduced after-hours referral rates by 77%.19 This reduction was attributed to physician’s determination that the problems did not require urgent referral, prior knowledge about the family’s ability to care for the patient, knowledge of the patients past medical history, and change in the patient’s condition.19 Interventional studies have been published with the goal of decreasing nonurgent PED usage. Grossman et al20 studied nonurgent Medicaid patients in the PED and utilized nurses and social workers to work with families for 3 months to assign them a PCP, eliminate barriers to care (such as transportation), and ensure follow-up. This intensive intervention resulted in 14% fewer nonurgent PED visits in the subsequent 6 months, but no effect at 12 months.20 Similar studies using telephone/written reminders to follow-up with the PCP following PED visits and reminders sent to pagers/email show short-term improved follow-up and decreased PED utilization in intervention groups, which is not sustained after 3 to 6 months.21-24 To our knowledge, no published studies have employed a collaborative PCP-specific strategy, which emphasizes follow-up appointments and instructs patients on the proper steps that should be taken in the future for similar nonurgent concerns. Importantly, our intervention is constructed with direct collaborative input from the PCP and is practice specific. We hypothesize that this PCP-specific intervention, with a relatively low cost and minimally labor intensive design, has the potential to have sustainable impact on future nonurgent PED usage.
Methods Study Design and Location This is a prospective randomized intervention trial conducted at the PED of a tertiary care children’s hospital. Study enrollment took place from February 2010 to May 2010 from 8 am to 8 pm, 5 days a week. Eligible children were 3 months to 16 years old whose caregiver selfidentified a PCP that was within a network of 35 cooperative practices. This practice group is composed of independent physician groups who have a central office for data sharing. Each practice is independently operated and has differing levels of capabilities to obtain labs/radiographs, variable hours, and different afterhours availability. At the time of the study, there were 35 physician offices in the group, comprising 185 pediatricians. Previous published studies have shown that these practices account for approximately 60% of the nonurgent visitations to the study PED.25
989 Further eligibility criteria for enrollment were the following: presentation with a primary caregiver responsible for medical decision making and nurse triage acuity as low acuity (defined as one of the lowest 2 triage categories on the 5-level Emergency Services Index triage system). The 2 lowest triage levels, indicate a likelihood for low resource utilization.16,26-28 Previous studies have used a validated method to determine visit urgency, which defined visits that resulted in a radiograph, laboratory study, electrocardiogram, electroencephalogram, or admission as urgent visits; all other visits were defined as nonurgent visits.26,29-31 Our study will define nonurgent patients as those with a low initial triage acuity who did not have a radiograph, laboratory study (other than rapid strep), electrocardiogram, electroencephalogram, and were not admitted. If initially triaged as a nonurgent visit based on triage acuity, but subsequently tests were ordered by the treating physician, the patient was no longer eligible to be in the study and was removed. Eligible patients were approached and consented for enrollment in the study prior to physician encounter. To ensure standardized consent and intervention sessions, study personnel received a 60-minute training session and followed a scripted description of the study. All consent forms and study methodology were approved by the institutional review board. All enrolled patients completed a questionnaire in the waiting room about what resources they used prior to coming to the PED for care and identified barriers to obtaining timely care. This questionnaire also assessed patient baseline satisfaction with access to their PCP office (Supplemental Information 1). The surveys were developed in accordance with previous guidelines32 and included a literature review of existing research, expert panel item generation, reduction, and pretesting. The survey used dichotomous and frequency-of-endorsement formats. Based on even or odd medical record numbers, a random sample of enrolled patients was selected to participate in the intervention arm. The intervention took place at the conclusion of the visit, after being seen by the physician and prior to discharge from the PED. Participants selected for the intervention were given a laminated one-page handout developed with input from their self-identified PCP which outlined office hours/location, scope of practice (ability to do radiographs, blood work, sutures, etc), and the preferred steps that patients should take to address medical concerns (Supplemental Information 2). Intervention subjects were told that the form was developed with direct input from their PCP. In a standardized 10-minute session, study representatives described the algorithm to caregivers and answered questions. Intervention patients were given a laminated copy of this one-page
990 form to take home. Control patients were managed with the standard of care; they were given routine discharge instructions by the PED nurse. Intervention patients similarly received standard discharge instructions. Spanish-speaking patients had a translated copy of the questionnaire and educational intervention form as required. If needed, an on call translator was used to answer questions in Spanish. The PCP-specific form used for intervention was developed in coordination with each of the participating 35 practices. A sample form was completed by the lead physician at each practice. If a practice used a telephone call center for after hours advice, the forms noted that a physician was on call every night and if the patient wished they could ask the call center to contact the afterhours physician. Each practice outlined what they felt were the optimal strategies to obtain this medical advice and how to best obtain same day sick visits. For the intervention and control groups, future visitation to either of the 2 local PEDs for both acute and nonacute visits was tracked via the electronic medical record at 6- and 12-month intervals. These are the only dedicated tertiary care PEDs in the area, so any returns to PEDs would be expected to occur at one of these sites. Since enrollment occurred over a period of several months, patients reached the 6- and 12-month intervals at slightly different times; this was accounted for in data extraction. In addition to PED visitation, future visits to the PCP office for both sick visits and well-child visits were quantified at the end of the 12-month study by examining billing data from the PCP offices. At 6 months after study enrollment was complete, patients in the intervention group were contacted by telephone to assess their satisfaction with access to their PCP office. Contact was attempted on 3 separate occasions; if the enrolling caretaker could not be contacted, the questionnaire was mailed in an addressed envelope. The questionnaire was identical to the one completed at study entry. Results were scored on a Likert-type scale and comparisons were made across the groups over the 6-month period. To be consistent with previous research, the timeliness composites, scored on a 4-point Likerttype scale, were dichotomized, with scores of >3.5 indicating the highest quality care and scores of
