News Magazine

Grady Health System’s (GA) Mobile Integrated Health Program

Photo provided by the authors.

A Statistical Analysis of Low-Acuity 911 Calls


Grady’s Mobile Integrated Health (MIH) program works to manage outpatient health concerns that otherwise burden emergency departments (Eds), improve quality of care and connect patients to the appropriate level of care and resources. This prospective study collected data from September 1, 2019 to March 31, 2020, to analyze Grady’s MIH response to low-acuity 911 calls compared to a traditional EMS (ACLS/BLS) response.

A total of 2,759 EMS calls were reviewed. These calls comprised the four most common emergency medical dispatch codes for Grady’s MIH response: i) “sick person other pain,” ii) “diabetic alert behaving normally,” iii) “back pain,” and iv) “falls.” Descriptive statistics and multivariable logistic regressions (MLR) were performed to compare disposition differences between MIH and traditional EMS services in whether calls were mitigated on-scene or transported. For MIH responses (n=300), 66.1{41490b4d0cf0dbc5ec3f65e11fff509c7d6ed2a53a838ebf7adf43f0908f07f3} were mitigated on-scene.


Comparatively, for traditional EMS responses (n=263), 11.4{41490b4d0cf0dbc5ec3f65e11fff509c7d6ed2a53a838ebf7adf43f0908f07f3} were mitigated on-scene. The MLR model found the odds that a patient was mitigated on-scene for an MIH response were 24 times that for an ACLS/BLS response (OR=24.19, p<0.001) after adjusting for patient sex, ethnicity, age, blood pressure, heart rate, pain response, glucose, time of day, and EMD code. The magnitude of the odds ratio significantly differed based on the dispatch code.

The results of this study indicate that utilizing Grady’s current MIH model is an effective way to mitigate low-acuity 911 concerns and decrease unnecessary ED utilization, while potentially reducing hospital readmissions and healthcare costs.


MIH is a rapidly evolving field that primarily uses EMS personnel, often in collaboration with other healthcare providers, to fill gaps in local healthcare infrastructure.1 The goals of these programs vary based on the challenges that prompted their creation; however, most goals include reducing unnecessary ED visits and EMS use, reducing hospital readmissions, and reducing healthcare costs.2 This emerging healthcare delivery model is producing promising clinical outcomes; however, there is a lack of reliable data to assess and compare these programs.3 MIH program evaluation data is needed for program planning to meet future health system needs.4

In response to increasing numbers of ED patients and EMS transports, Grady Health System’s EMS division in Atlanta, GA developed and implemented an MIH unit in January 2013.5 Grady’s EMS system is the hospital-based emergency care component of Grady Health System that provides 911 services to the city of Atlanta, East Point, and College Park, GA (6). Initially started as an Upstream Crisis Intervention Unit focused on behavioral health, Grady’s MIH program evolved in the last several years as the critical needs of the Atlanta community emerged.7, 8

Currently, the global aim of Grady’s MIH program is to enhance community-based healthcare, manage outpatient health concerns that otherwise burden EDs, improve quality of care, reduce healthcare costs, and connect patients to the appropriate level of care and resources.5 The program has several initiatives to address improving 911 system management of non-emergent calls and high ED utilizers, including reducing 30-day readmissions and mitigating gaps between inpatient and outpatient care.9

Each functional unit of Grady’s MIH program consists of a nurse practitioner or physician assistant partnered with an advanced emergency medical technician (AEMT) or paramedic. MIH is relatively new to the discipline of medicine and nursing. As such, most nurse practitioner and physician assistant training programs do not include EMS-related topics and few advanced practice providers have experience working within an EMS system.

The new-hire training for Grady MIH providers is conducted over a period of 14 days that includes four learning objective areas: dispatch, 911 response, chronic disease management, and community resources. Providers are immersed in the 911-call center to receive instruction on the dispatch system, call assignment, radio etiquette, and field safety protocols. The nurse practitioners and physician assistants complete shifts on an ambulance for familiarization with call response, emergency equipment usage, and 911 documentation.

These providers are also familiarized with community and hospital-based resources to assist with the needs that impact utilization and health outcomes. Finally, a skills lab was also created to reinforce training on EKG interpretation, wound and ostomy care, and skin closure techniques.

These units deploy within Grady’s 911 system in Fulton County, Georgia.9 Each unit responds in a fully equipped medical first responder vehicle with the capabilities to provide high-level care including 12-lead electrocardiograms, medications, airway management, and point-of-care testing.9

These units are activated for calls based on the same 911 triage system and emergency medical dispatch (EMD) code system as Grady’s traditional EMS response. There are 66 EMD codes that can be assigned a MIH response. These calls cover non-emergent chief complaints that often do not require evaluation or management in an acute care setting. In addition to the on-scene providers, Grady’s MIH program has a Nurse Navigator who connects MIH patients with community resources, including transportation, appointment scheduling, and medication refills.

The purpose of this study was to analyze Grady’s MIH response to low-acuity 911 calls compared to traditional EMS (ACLS/BLS) units. This analysis was designed to build the evidence base to support the growth of other MIH EMS programs and contribute to national development and expansion of this innovative model of care.


Data Collection and Variables

The study population consisted of all patients in Atlanta and Fulton County, GA, who called 911, from September 1, 2019 to March 31, 2020, with the top four most common low-acuity complaints that qualified for an MIH response. De-identified data were electronically extracted from Grady’s EMS charting system used to document 911 patient encounters. Prior to data extraction and analysis, this study received IRB exemption from Emory University in Atlanta, Georgia.

The following variables were collected for the data analysis: EMS unit number, call outcome, time of call, patient age, sex, ethnicity, systolic and diastolic blood pressure, pulse, respiration rate, temperature, blood glucose, patient-reported pain (measured on a 0-10 scale), and the EMD code. These variables were collected for the top four most common MIH EMD codes, which were: i) “sick person other pain,” ii) “diabetic alert behaving normally,” iii) “back pain,” and iv) “falls.”

All of these are determined by dispatch during the 911 call prior to dispatching an EMS or MIH unit. For “sick person other pain,” the patient is determined to be alert, breathing normally, not bleeding, and does not have chest pain, sickle cell crisis, thalassemia, or any other priority symptoms. For “diabetic alert behaving normally,” the patient is completely alert, breathing normally and behaving normally.

For “back pain,” the patient is alert, breathing normally and did not have a recent fall or trauma, no aortic aneurysm, and no fainting or near fainting. For “falls,” the patient is alert and breathing normally, the fall was a ground level fall, there is no serious bleeding, the area of the body injured is not dangerous (i.e. not chest or neck) and there are no obvious deformities.

Statistical Analysis

The exposure of interest, MIH vs ACLS/BLS, was determined from the EMS unit number. Call outcome, the outcome of interest, was recoded into two variables: mitigated on-scene versus transported. A mitigated patient was defined as a patient who was examined and remained on-scene with no transport required at the end of the encounter. A transported patient was defined as a patient who was examined and then moved to any other destination, such as a hospital or clinic.

This would include all patients with urgent and emergent complaints. Call time was dichotomized into MIH operating hours (7 a.m.-10 p.m.) and non-MIH operating hours (10 p.m.-7 a.m.). Patient vital signs, reported pain and blood glucose were averaged across two separate readings taken during the EMS response. Vital signs and blood glucose were dichotomized into normal and abnormal. The normal ranges for vital signs were determined using the Cleveland Clinic recommendations.10 The normal range for blood glucose was determined using the Current Medical Diagnosis and Treatment textbook.11 Ordinal pain scores were recoded into two groups using the median as the cutpoint (0-5 and 6-10).

Descriptive statistics including t-tests, chi-squared test, and Fisher’s exact tests were performed to compare MIH vs ACLS/BLS, when appropriate for the variables of interest. The primary multivariable logistic regression model (MLR) estimated the association that MIH versus ACLS/BLS had with the call outcome (i.e., the patient being mitigated or transported) adjusted for age, sex, ethnicity, systolic blood pressure, diastolic blood pressure, heart rate, glucose, pain, EMD code and hours of operation.

The model covariates were chosen based on their accessibility, clinical significance, and potential predictive value to the call outcome. After interpreting the primary MLR model, four sub-analyses were additionally conducted restricting the dataset to each of the EMD codes. SPSS® Statistics version 26 was used for the analysis.


Out of the 2,759 EMS calls, 83.5{41490b4d0cf0dbc5ec3f65e11fff509c7d6ed2a53a838ebf7adf43f0908f07f3} were responded to by a traditional EMS unit (ACLS/BLS unit), and 16.5{41490b4d0cf0dbc5ec3f65e11fff509c7d6ed2a53a838ebf7adf43f0908f07f3} were responded to by an MIH unit. For both MIH and ACLS/BLS units, the most common EMD code was “sick person other pain,” followed by “back pain,” “diabetic alert behaving normally” and “falls.”

For MIH responses, 66.1{41490b4d0cf0dbc5ec3f65e11fff509c7d6ed2a53a838ebf7adf43f0908f07f3} were mitigated on-scene, and 33.9{41490b4d0cf0dbc5ec3f65e11fff509c7d6ed2a53a838ebf7adf43f0908f07f3} were transported (Table 1). Comparatively, for ACLS/BLS responses, 11.4{41490b4d0cf0dbc5ec3f65e11fff509c7d6ed2a53a838ebf7adf43f0908f07f3} were mitigated on-scene, and 88.6{41490b4d0cf0dbc5ec3f65e11fff509c7d6ed2a53a838ebf7adf43f0908f07f3} were transported. The mean age of MIH patients was 51 years, compared to 56 years for ACLS/BLS patients which was statistically different (t-test p=0.04). Approximately 49{41490b4d0cf0dbc5ec3f65e11fff509c7d6ed2a53a838ebf7adf43f0908f07f3} of the patients were female in MIH response compared to 51{41490b4d0cf0dbc5ec3f65e11fff509c7d6ed2a53a838ebf7adf43f0908f07f3} in ACLS/BLS response which was not statistically different (chi-squared p=0.53).

Similarly, there were no statistically significant differences between the groups in ethnicity comparisons (Fisher’s exact test p=0.341). For MIH patients, 71{41490b4d0cf0dbc5ec3f65e11fff509c7d6ed2a53a838ebf7adf43f0908f07f3} had abnormal systolic blood pressure and 45{41490b4d0cf0dbc5ec3f65e11fff509c7d6ed2a53a838ebf7adf43f0908f07f3} had abnormal diastolic blood pressure while ACLS/BLS patients 81{41490b4d0cf0dbc5ec3f65e11fff509c7d6ed2a53a838ebf7adf43f0908f07f3} had abnormal systolic blood pressure and 60{41490b4d0cf0dbc5ec3f65e11fff509c7d6ed2a53a838ebf7adf43f0908f07f3} had abnormal diastolic blood pressure. For MIH patients, 69{41490b4d0cf0dbc5ec3f65e11fff509c7d6ed2a53a838ebf7adf43f0908f07f3} of patients had abnormal glucose and 63{41490b4d0cf0dbc5ec3f65e11fff509c7d6ed2a53a838ebf7adf43f0908f07f3} of ACLS/BLS patients had abnormal glucose.

For MIH patients, 63{41490b4d0cf0dbc5ec3f65e11fff509c7d6ed2a53a838ebf7adf43f0908f07f3} had high pain and 56{41490b4d0cf0dbc5ec3f65e11fff509c7d6ed2a53a838ebf7adf43f0908f07f3} of ACLS/BLS patients had high pain. Systolic blood pressure, diastolic blood pressure, glucose groups, and pain scale groups were all statistically different between MIH and ACLS/BLS (chi-squared p<0.001, p<0.001, p=0.029, p=0.010 respectively). For both MIH and ACLS/BLS, 20{41490b4d0cf0dbc5ec3f65e11fff509c7d6ed2a53a838ebf7adf43f0908f07f3} of patients had an abnormal heart rate (chi-squared p=0.948).

The MLR model found the odds that a patient was mitigated on-scene for an MIH response were 24 times that for an ACLS/BLS response (OR= 24.19 p<0.001) (Table 2). This result was independent of patient sex, ethnicity, age, blood pressure, heart rate, pain response, glucose, time of day and EMD code. 

Sub-analyses of the primary MLR model were conducted restricting the dataset to each of the four EMD codes (Table 2). In all four models, patients were more likely to be mitigated on-scene with an MIH response compared to an ACLS/BLS response. However, the estimated MIH versus ACLS/BLS odds ratio differed across the four models. Specifically, among patients who were triaged with the EMD code of “sick person other pain,” the odds of being mitigated on-scene for an MIH response were 27 times that of ACLS/BLS response (OR= 27.09, p< 0.001).

Among patients who were triaged as “diabetic alert behaving normally,” the respective odds ratio was 15 (OR=15.34, p<0.001). Among patients who were triaged as “back pain,” the respective odds ratio was 196 (OR=196.10, p<0.001). Among patients who were triaged as “falls,” the respective odds ratio was 4 (OR= 4.19, p= 0.003). The interaction between units and determinants was tested and determined to be statistically significant (chi-square value 36.706 under 3 degrees of freedom, p<0.001).


This study was designed to analyze Grady’s MIH response to low-acuity 911 calls through data collection and statistical analysis. We found that MIH units mitigate patients on-scene more often than ACLS/BLS units for the four most common EMD codes even while controlling for numerous key variables, including demographics, vital signs, and hours of operation.

This study demonstrates that Grady’s current MIH model effectively mitigates low-acuity 911 responses, thereby decreasing inappropriate ED utilization. These findings are consistent with Gregg et al.’s systematic review, which states that MIH programs result in significant decreases in transports to the ED and ED utilization.12

By controlling for patient demographics, type of call, patient acuity, and time of operation, the results suggest that the advanced practice provider’s involvement within the EMS response is a likely reason these patients were mitigated on-scene. This finding is consistent with prior literature that suggests that these providers are critical to the success of MIH programs.13 Increasing the utilization of nurse practitioners and physician assistants in MIH helps achieve increased on-scene mitigation, decreased ED utilization, and additional cost savings for both patients and healthcare systems. 

Of particular interest is the difference in the likelihood of on-scene mitigation comparing the four EMD codes. Of the four EMD codes, patients are most likely to be mitigated on-scene by MIH for “back pain” followed by “sick person other pain,” “diabetic alert behaving normally,” and “falls.” Since “falls” patients were the least likely to be mitigated on-scene by MIH compared to the other three codes, further investigation could demonstrate if a change in the triage process related to dispatching MIH for falls is needed to improve the use of MIH resources.

One limitation is that this study only used data from ESO, the electronic charting system used by AEMTs and not data from Epic®, the charting system that the advanced practice providers utilize. The charting performed by the advanced practice providers is significantly more robust than the charting performed by the AEMTs and could result in further data findings.

Another data limitation was the inability to include all vital signs in the regression models; specifically, there were not enough data points in the abnormal respiratory rate and temperature groups to include these important variables in the MLR models. A race variable may provide more insight than just using an ethnicity variable.

Unfortunately, race was not readily available in our dataset. Another limitation is that patients in the dataset can be represented multiple times if they called 911 with a low-acuity complaint more than once during the study period. A patient being represented multiple times could bias the results as patient health characteristics can be correlated from one episode to another rather than every encounter being independent.

An additional factor not considered in our analysis is the novel coronavirus. Given that our collection period did not overlap with the vast increase in EMS calls due to COVID-19, we do not anticipate that factor to bias our results. However, we recognize the importance of further research of the impact of COVID-19 on MIH versus ACLS/BLS scene mitigation and transfer rates.

Additionally, the study did not look at subsequent data to identify if any of the mitigated calls still required transport after the MIH encounter. Grady EMS has a robust Quality Assurance team and has not reported any concerns regarding inappropriate MIH mitigations that later resulted in poor patient outcomes.

This study’s findings support that patients with the same four low-acuity 911 complaints are more likely to be mitigated on-scene with an MIH response compared to an ACLS/BLS response. Utilizing similar analytic methods to determine which 911 calls are most likely to result in on-scene mitigation could help other systems tailor MIH programs to their current 911 needs.

This would be particularly helpful in resource-limited settings where there may be few MIH providers and MIH units need to be used as wisely as possible. Implementing Grady’s MIH model offers the unique opportunity to potentially address social determinants of health, bridge the gap between outpatient medicine and emergency medicine, and improve care outcomes at decreased costs. 


  1. Choi BY, Blumberg C, Williams K. Mobile integrated health care and community paramedicine: An emerging emergency medical services concept. Ann Emerg Med. 2016 Mar 1;67(3):361–6.
  2. Patterson DG, Coulthard C, Garberson LA, Wingrove G, Larson EH. What is the potential of community paramedicine to fill rural health care gaps? J Health Care Poor Underserved. 2016;27(4A):144–58.
  3. Rasku T, Kaunonen M, Thyer E, Paavilainen E, Joronen K. The core components of community paramedicine – integrated care in primary care setting: A scoping review. Scand J Caring Sci. 2019 Sep;33(3):508–21.
  4. Wilcoxson WJ. Community paramedicine and mobile integrated health care: Existing resources bringing new benefits. South Med J. 2016 Mar;109(3):151–3.
  5. Grady Emergency Medical Services. Mobile integrated health [Internet]. 2018 [cited 2020 Feb 20]. Available from:
  6. Grady Emergency Medical Services. About Grady EMS [Internet]. Grady Health. 2020 [cited 2020 Feb 12]. Available from:
  7. Colman M. Clinician-embedded MIH program improves EMS use in high-risk population [Internet]. EMS1. 2017. Available from:
  8. Grady Emergency Medical Services. Upstream crisis intervention unit. Atlanta, GA; 2013 May.
  9. Grady Emergency Medical Services. MIH abstract. Atlanta, GA; 2019.
  10. Cleveland Clinic. Vital Signs [Internet]. Cleveland Clinic. 2020 [cited 2020 Jul 21]. Available from:
  11. Papadakis MA, McPhee SJ, Rabow MW, editors. Current medical diagnosis & treatment 2020. 59th ed. McGraw-Hill Education; 2020.
  12. Gregg A, Tutek J, Leatherwood MD, Crawford W, Friend R, Crowther M, et al. Systematic review of community paramedicine and EMS mobile integrated health care interventions in the United States. Popul Health Manag. 2019 Jan 3;22(3):213–22.
  13. Georgiev R, Stryckman B, Velez R. The integral role of nurse practitioners in community paramedicine. J Nurse Pract. 2019 Nov;15(10):725–31.