Cost–benefit analysis of telehealth in pre-hospital care

Study design

This study utilized a case-control research design to evaluate the City of Houston Emergency Medical Services initiative called Emergency Telehealth and Navigation (ETHAN). ²⁰ The study intervention group incorporated telehealth to screen primary care patients and then navigate eligible patients towards non-ED locations using taxi-based transport instead of more resource-intensive ambulances. In this community surrounding Houston, there is only one primary EMS responder that accepts emergency 911 calls—the Houston Fire Department. All other EMS agencies support smaller surrounding communities, and private EMS companies are solely intended for inter-facility transfers. These locations were community health clinics or hospital-based primary care clinics, which could provide more appropriate levels of care. We calculated a necessary sample size of 5000 for the intervention group.

We identified retrospectively a matching control group, of similar patients that did not receive the telehealth-based intervention. A matching algorithm identified similar primary chief complaints related patients based on the individual factors of age, patient chief complaint, gender, and geography. The control group was traditional ‘treat and transport’ to local hospital ED. We then compared the control group costs and outcomes against those enrolled in the telehealth intervention using time, costs, and clinical data captured in the electronic patient care record system (PCR).

Intervention protocol

Enrolment in the study was initiated by EMS personnel. When the first responding EMS apparatus arrived on scene, the crew assessed the patient and determined whether the patients were in need of urgent or primary care. If the patient was considered non-acute or urgent, and they met the inclusion criteria, the paramedic crews chose whether to initiate the telehealth intervention. All of the EMS units in Houston were equipped with mobile tablet-based technology installed with HIPAA-compliant and secure video teleconferencing software that would enable consult between the patient and an emergency physician residing in EMS headquarters. On the other end of the telehealth consultation, residing in the ‘base station’ at the emergency centre from approximately 7 am to 7 pm six days per week, were board-certified emergency medicine physicians. There are 16 part-time, contracted EMS physicians in total with only one physician on-staff at any point in time.

Patient inclusion criteria for the ETHAN intervention included only patients who were determined to be primary care through the primary assessment performed by the responding EMS crew. Patients had to have no acute, urgent or life-threatening conditions that would require immediate treatment and transport. In addition, patients had to be able to communicate with the EMS physician via the tablet and present with vital signs that were age-appropriate and within normal ranges.

The information taken by the EMS crew at the scene was recorded in an electronic PCR on the tablet, which also contained the telehealth solution. The connected PCR enabled the emergency physician to access the patient’s records created at the scene, including demographics, vital signs, medical history, allergies, medications, and chief complaint. While the patient consulted with the physician via teleconferencing, the field crew supported the physician with any additional needs including taking a new set of vital signs. The physician, based on their medical expertise, determined the disposition of the patient, which included one of the following:

transport to hospital ED via ambulance;

The primary aim of the study was to navigate primary care-related patients away from the ED where possible, and towards the primary care clinics. It was hypothesized that costs would be lower if this aim were achieved. Since patients had to agree to this disposition, the third option allowed for a taxi alternative rather than ambulance transport.

Data analysis

We conducted a cost–benefit analysis of a telehealth intervention in a community setting. Direct costs were calculated for resource utilization, measured as average actual time spent by both the EMS field crews and physicians during the telehealth consultation, computed ambulance and vehicle costs for transports, and costs for equipment and technology for startup and operations. We measured both direct and indirect costs for the computing hardware and software costs involved in the implementation of telehealth model. Other indirect costs included the cost of programme administration. All other overhead costs (including administrative salaries, space, and utilities) were the same in both groups and were excluded from the analyses.

Major activity cost for the control and the intervention mode were estimated by a combination of prospective micro- and gross-costing. Gross-costing was used for the capital investments (e.g. vehicles, transportation, technology) and micro-costing was used to estimate all time and resource utilization for personnel and operations.

The largest portion of the provider costs for both groups was EMS personnel time. We calculated this by measuring the start and stop time for personnel on each incident, measured as time of call from patient to end of care for the patient. We multiplied this by the number of personnel on scene and the average actual hourly personnel cost per minute, derived from payroll sources. Similarly, we computed direct time estimates for the physicians involved in either a telehealth consultation.

Transportation costs included vehicle acquisition costs, fuel, maintenance and repair, as well as the costs for alternative taxi transportation and were amortized by the number of incidents and average duration of that incident. Technology costs included the initial and ongoing telemedicine investments, plus computer and tablet hardware.

Financial benefits included averted costs as well as the change in the relative number of ED visits (% ED utilization). From the hospital ED perspective, reduction in ED utilization was a primary goal and benefit of this programme. This included both the number of patients transported by ambulance and taxi transports to the ED. To measure the cost-benefit of this change in ED visits, we used the product of the absolute change in ED visits multiplied by the primary care-related ED (PCRED) cost rate ($950) for Houston Texas. This cost estimate was obtained from a local health economics academic study that determined the PCRED costs for the urban hospitals at an average of $780/PCRED visit. ⁴ We adjusted this for five years at an average 4% healthcare inflation rate (from the Bureau of Labor Statistics) to obtain a $950 cost estimate in current dollars.

Cost–benefit was defined as the cost averted benefit resulting from reduction in ED visits and transports (measured in 2015 USD). In addition to the provider perspective, we calculated the societal cost savings as well, which included both the cost of ED visits averted from non-transport and the change in provider costs. To normalize the savings, we compared the net change in costs against the net change in benefits, to get a cost per ED visit avoided.

SPSS was used to perform data analysis (SPSS Statistics, version 24, IBM Corp, Armonk, NY). The significance level of α = 0.05 was applied. Differences were compared using non-parametric tests (Mann–Whitney U test) and parametric tests (Student’s t test) based on the distribution of the variables.