The availability of telecardiology consultations and transfer patterns from a remote neonatal intensive care unit

Abstract

We examined records of all admissions to an isolated community neonatal intensive care unit (NICU) in California between 2001 and 2006. We also reviewed the echocardiograms for diagnosis, disposition of patient and necessity for transport. In 2004, a telemedicine link (mainly store-and-forward) was established to a university children's hospital (UCH) 290 km away. The number of NICU patients having an echocardiogram increased from 280 (27% of 1029 admissions) to 385 (40% of 963, P = <0.001) after telemedicine became available. There was an increase in the proportion of normal studies, from 31% to 37% (P = 0.03), and an increase in the number of patients diagnosed with cardiac pathology from 192 (19% of all admissions) to 241 (25%, P < 0.001). Twenty-four patients were transferred for cardiac reasons during each three-year period; however seven pre-telemedicine transfers were avoidable, compared with two post-telemedicine transfers (P = 0.06). There was a change in referral pattern (65% to the UCH pre-telemedicine, compared with 78% post-telemedicine) although it was not significant (P = 0.10). Thus the availability of the telecardiology link was associated with increases in the utilization of echocardiography, in the proportion of normal studies, and in the percentage of neonates diagnosed with cardiac pathology without an increase in the number transferred for cardiac reasons. There was a reduction in unnecessary transfers and a strengthened relationship with the centre providing the telecardiology service.

Introduction

The incidence of congenital heart disease needing intervention in infancy is about 3 per 1000 births,¹ and a large proportion of these patients will be born in rural and remote medical centres where paediatric cardiology expertise is not available. Definitively excluding cardiac disease is a challenge for a neonatologist or paediatrician practising in centres where an experienced sonographer and paediatric cardiologist are not available.

The most common way of addressing this lack of access is to use store-and-forward echocardiography, i.e. an adult sonographer performs an echocardiogram in the nursery or neonatal intensive care unit (NICU) and the recording on a videocassette is mailed to a paediatric cardiologist at a tertiary care centre. However, because of the time this takes, life-saving interventions are sometimes delayed. As technology has improved, faster ways of sending diagnostic imaging studies have been developed.^2,3 By the late 1990s, paediatric cardiologists could remotely evaluate NICU patients in real-time.^4–6 More recently, it has become possible to transmit complete digital echocardiograms rapidly over high bandwidth connections, thus reducing the turnaround time for store-and-forward echocardiography.^7–9

Success with real-time telemedicine echocardiography has been reported in numerous settings and countries.^6,10–17 However, there appear to be no reports of the successful use on a large scale of digital store-and-forward echocardiography. Furthermore, while the economic impact of telemedicine has been estimated in a number of previous studies, only one study has looked retrospectively at the pre- and post-telemedicine (using real-time echocardiogram transmission) periods to examine changes in patient care and referral patterns.⁹

Telecardiology

In 2004, both real-time and store-and-forward telecardiology was established between a university-affiliated children's hospital (UCH, the University of California, Davis Children's Hospital) in Sacramento and a remote community hospital (RH, the Mercy Medical Center Redding) located 290 km away. The RH is located in a town with a population of 90,000 and approximately 2100 live births per year, and is the only hospital in the area with a level III NICU. The RH NICU treats patients in need of mechanical ventilation (including high frequency ventilation), but not inhaled nitric oxide or extracorporeal membrane oxygenation (ECMO). With the exception of three neonatologists, no paediatric subspecialists are available there.

The telemedicine connection was made with a T-1 connection (1.5 Mbit/s). Videoconferencing equipment (Tandberg 880, Tandberg, Norway) was installed at the RH NICU, several locations at the UCH, and in the home of one of the UCH paediatric cardiologists. For echocardiograms, an adult sonographer would perform the study and transmit the images to a server at the UCH (KinetDx, Siemens, Mountain View, California). The paediatric cardiologist at the UCH would then be paged to report the study. Treatment recommendations were made via telephone conversation between the UCH cardiologist and RH neonatologist.

If a real-time consultation was deemed necessary by the referring doctor, the cardiologist could be connected to the telemedicine link, either at UCH or at home. The real-time telemedicine link enabled the cardiologist to see the neonatologist and the patient, and view the images being displayed on the ultrasound machine, enabling the sonographer to perform additional imaging as required while guided verbally by the cardiologist. Once the evaluation was complete, recommendations for medical care or transfer were made and communicated via the videoconference to both the neonatologist and the family of the newborn.

Aims

The purpose of the present study was to evaluate differences in echocardiography utilization and patterns of transfer before and after the implementation of a primarily store-and-forward echocardiography telecardiology programme in a remote NICU without any other access to a paediatric cardiologist. We also hypothesized that the programme would reduce the number of inappropriate transfers.

Methods

We compiled a list of all the echocardiograms that were completed in the three years prior to the start of the telecardiology programme as well as the first three years after the programme began. The echocardiograms were divided into the categories of normal anatomy, simple or complex congenital heart disease, patent ductus arteriosus (PDA) and persistent pulmonary hypertension of the newborn (PPHN). Patent foramen ovale and atrial shunts in the region of the foramen ovale less than 3 mm in diameter were considered normal. PDA was considered abnormal given that the standard at the RH generally was not to perform echocardiography for screening of asymptomatic infants.

To assess the accuracy of the initial echocardiogram reading, we also reviewed the results of any follow-up studies; when more than one echocardiogram was performed on a single patient, only the results of the first echocardiogram were recorded for the purposes of this study, i.e. each baby represented a data point, not each echocardiogram.

We evaluated the records of all the patients who were transferred from the RH (for any reason), reasons for transfer (diagnosis), and (for purposes of determining necessity of transfer) clinical condition at transfer including oxygen saturation, ventilator support and inotropic medications being administered. For the patients who were transferred, the hospital course at the tertiary care centres (including treatment with inhaled nitric oxide, ECMO, need for surgical or medical intervention for PDA, discharge disposition), final cardiac diagnosis if any, and any surgical and transcatheter interventions required were noted.

The transfers for cardiac reasons (simple or complex congenital heart disease, PDA or PPHN) were then further divided into two categories – avoidable and unavoidable. Any neonate who was transferred from the RH with a concern for congenital heart disease, but who ultimately did not have congenital heart disease in need of neonatal surgical repair or catheter intervention, was considered to have been transferred unnecessarily (the echocardiogram was reviewed by one of the authors in all of these cases). Neonates who were transferred for surgical ligation of PDA, who did not undergo ligation because of ductal closure documented at the receiving hospital, were considered to be transferred unnecessarily. Neonates transferred with a diagnosis of PPHN were considered inappropriately transferred if their clinical status was improved on arrival at the tertiary care centre, as evidenced by decreased pressor use, improved oxygenation, improvement on echocardiogram, and if there was lack of further escalation of support, use of inhaled nitric oxide, or use of extracorporeal life support.

The study was approved by the appropriate ethics committees.

Results

The total number of live births at the RH increased over the study period by an average of 3.3% per year (Table 1). The percentage of live births admitted to the NICU remained relatively constant at 17% (range 15–19%). The number of patients having at least one echocardiogram was 280 (27% of admissions) in the pre-telemedicine period and increased to 385 (40% of admissions) (P < 0.001) when telemedicine was available. There was also a significant increase in the percentage of normal first echocardiograms (Figure 1): 88 of the 280 studies (31%) were normal from 2001–2003, and 144 of the 385 studies (37%) from 2004–2006 (P = 0.03).

Figure 1

Neonates undergoing echocardiograms, by year. Normal echocardiograms displayed in white, abnormal in grey. The proportion of abnormal echocardiograms (exclusive of patent foramen ovale and atrial shunt less than 3 mm in diameter) was lower (P = 0.03) after telemedicine began in 2004 than in the 3-year period before. However the difference between the single year prior to telemedicine and any of the subsequent years was not

Table 1

Birth, NICU admissions and transfer data and echocardiography utilization at the remote hospital

	Pre-telemedicine			Post-telemedicine
	2001	2002	2003	2004	2005	2006
Live births (n)	1844	1913	2051	1994	2066	2163
NICU admissions (n)	319	319	391	295	329	339
Neonates ≥ 1 echocardiogram (% of NICU admissions)*	56 (18%)	92 (29%)	132 (34%)	118 (40%)	135 (41%)	132 (39%)
Normal echocardiograms	10	30	48	33	61	50
Abnormal echocardiograms**	46	62	84	85	74	82
NICU admissions with an abnormal echocardiogram (%)*	14%	19%	21%	29%	22%	22%
Transported from the RH (% of all NICU admissions)^#	16 (5%)	20 (6.3%)	24 (6.1%)	18 (6.1%)	21 (6.4%)	28 (8.3%)
Cardiac reason for transport	8	6	10	6	13	5
Non-cardiac reason for transport	8	14	14	12	8	23

*Significant increase, 2004–2006 versus 2001–2003 (P < 0.001)

**Simple or complex congenital heart disease, patent ductus arteriosus, persistent pulmonary hypertension; patent foramen ovale and atrial level shunt less than 3 mm in diameter are not included in the ‘abnormal’ category

^# P = 0.33, 2004–2006 versus 2001–2003

In the pre-telemedicine period (2001–2003), 280 patients were evaluated with echocardiograms. The most common diagnosis was PDA (n = 120, 43% of all echocardiograms) (Table 2). After the telecardiology programme was implemented, 385 patients had echocardiograms (all store-and-forward as the initial study). The case-mix was similar, except that the proportion of normal cases was slightly higher (P = 0.03) after the introduction of telemedicine (Table 2).

Table 2

Echocardiographic diagnoses during the study period

	Pre-telemedicine	Post-telemedicine
Patients with echocardiograms (n)	280	385
PDA	120 (43%)	156 (41%)
Normal*	88 (31%)	144 (37%)
PPHN	14 (5%)	12 (3%)
Poor function	2 (1%)	2 (1%)
Congenital heart disease	56 (20%)	71 (18%)
Isolated ASD or VSD	43	50
Left heart obstruction	5	8
Cyanotic	2	6
Other	6	7

ASD = atrial septal defect; PDA = patent ductus arteriosus; PPHN = persistent pulmonary hypertension of the neonate; VSD = ventricular septal defect. ‘Other’ included simple valve abnormalities, tumours, hypertrophic myopathies

*P = 0.03

Initial impressions were confirmed in all cases, and there were no diagnostic errors. Prior to 2006 accurate records pertaining to additional real-time consultation and imaging were not kept: in the last year of the study these consultations were documented in the records of both institutions, with 19 patients (14%) having real-time consultation in addition to their store-and-forward echocardiogram.

A total of 127 patients was transferred from the RH between 2001 and 2006 (Table 3). The overall rate of transfer (5.8% versus 7.0%) did not change between the two periods (P = 0.33). Forty-eight of the transfers were for possible cardiac disease, PDA or PPHN. The other 79 patients were transferred for social reasons or need for clinical expertise in a specialty unavailable at the RH. There were no deaths at the RH attributable to cardiac disease and no pre-procedure deaths in infants transferred for cardiac intervention during the study period.

Table 3

Transfer destination (disposition) of infants from the remote hospital during the pre- and post-telemedicine periods*

	Pre-telemedicine	Post-telemedicine
To the UCH	39 (65%)	52 (78%)
For cardiac reason	15	20
For non-cardiac reason	24	32
To other hospitals	21 (35%)	15 (22%)
For cardiac reason	9	4
For non-cardiac reason	12	11

*Numbers represent actual patients, with percent of total transfers in the period in parentheses. Cardiac reasons (as defined in the text) included transfer diagnosis of simple or complex congenital heart disease, patent ductus arteriosus in need of surgical ligation, or persistent pulmonary hypertension

Pre-telemedicine transfers

In the period before telemedicine was available (2001–2003), there were 24 transfers for possible cardiac disease including PDA or PPHN. The most common reason for transfer was for a possible need for ECMO and inhaled nitric oxide for patients with PPHN (n = 9) (Table 4). Total transfers to the UCH represented 65% of the total number of transfers from the RH to tertiary care institutions.

Table 4

Clinical indication for neonatal transfer from the remote hospital to tertiary centres during the study period

	Pre-telemedicine	Post-telemedicine
PDA	7	6
PPHN	9	5
Arrhythmia	0	2
Congenital heart disease	8	11
Atrioventricular septal defect	1	1
Left heart obstruction	4	5
Cyanotic	1	3
Tumours, myopathies, other	2	2
Non-cardiac	36	43
Total transfers (% of all admissions)*	60 (5.8%)	67 (7.0%)

PDA = patent ductus arteriosus; PPHN = persistent pulmonary hypertension of the neonate

*P = 0.33, 2004–2006 versus 2001–2003

Of the 24 patients transferred for cardiac reasons, seven were determined to be avoidable. Two patients had mild PPHN, one had acyanotic tetralogy of Fallot, one had a PDA which required no treatment, and one haemodynamically stable patient with non-obstructive cardiac rhabdomyomas (in association with tuberous sclerosis) was transferred urgently, presumably due to a lack of familiarity with this diagnosis on the part of the sonographer.

Of the 17 patients whose transfers were unavoidable, there was one patient worthy of mention. Although preliminary findings by echocardiogram were of PDA, following review of the videotape of the echocardiogram by a cardiologist on its arrival at UCH two days later, critical coarctation was diagnosed. The neonate was immediately started on prostaglandin treatment and transferred to a tertiary care centre.

Post-telemedicine transfers

In the period after the telemedicine programme was established, there were 24 transfers for possible cardiac disease. The most common reason for transfer was congenital heart disease (n = 11) (Table 4). Total transfers to UCH represented 78% of the total from the RH.

Only two of the cardiac transfers were deemed avoidable. One was a patient with a PDA transferred for surgical ligation: at the tertiary care centre, the patient was given indomethacin, which resulted in ductal closure. The second patient was transferred because of a suspected hypertrophic cardiomyopathy. At the tertiary care centre, hypertrophy due to twin-twin transfusion syndrome was diagnosed. This patient did not receive any cardiac intervention and was eventually transferred back to the RH.

Discussion

We found that the availability of telecardiology was associated with fewer inappropriate transfers for cardiac reasons (7 out of 24 prior to telemedicine versus 2 out of 24 after implementation). The overall rate of transfer did not change significantly during the two study periods. There were slightly fewer transfers for cardiac reasons in patients who had abnormal echocardiograms: 12.6% (24/191) prior to telemedicine versus 10.3% (24/242) after implementation, despite an increase in the percentage of neonates with abnormal echocardiograms: 19% (n = 192) versus 25% (n = 241) of admissions, P < 0.001. The transport destination was more likely to be within the UCH system in the telemedicine period than it was in the period prior, for both cardiac and non-cardiac transfers (Table 3).

Previous studies have evaluated neonatal telecardiology and the potential for a reduction in unnecessary transfers. Randolph et al. ¹⁵ examined the use of real-time telecardiology and found that an ‘immediate change in local medical management’ was made in 19% of the 133 cases that they reviewed. Seven transfers were avoided altogether, and although they speculated that telemedicine could decrease the overall number of transfers, they had no control group for comparison. Sable et al. ⁶ found that the use of telemedicine echocardiograms in rural Louisiana led to changes in medical management in 42% of 60 cases and stated that at least five transfers were avoided. Finley et al. ¹¹ published similar results in Canada at about the same time. Other studies have shown the effectiveness of telemedicine in other locations.^{10,12–14,16,18} All of these prior studies have lacked a control group for comparison of pre- and post-implementation rates of NICU transfers. Only one other study has compared the rates of, reasons for, and patterns of transfer before and after the implementation of a telecardiology system.⁹

From the perspective of transport utilization, the findings in our study were similar to these previous studies: seven transfers could have been avoided altogether in the pre-telemedicine period compared to two in the post-telemedicine period. In addition, the potential for improved care because of the more rapid transport possible with telemedicine availability was particularly apparent in our study for the patient whose critical coarctation of the aorta was initially missed on the preliminary echocardiogram in the pre-telemedicine era. There were no such diagnostic errors in the post-telemedicine era.

Telemedicine has potential benefits beyond the improvement in medical management. For the tertiary referral hospital, utilization of cardiology and NICU services may increase, as suggested by our observation that a higher percentage of the RH transfers were sent to the UCH after telecardiology was available. For the rural or isolated community hospital, telemedicine may assist in patient retention, enabling the hospital to keep patients without surgical heart disease or PPHN in their NICU, and allowing patients and their families to remain in the local area rather than face the often considerable burden of travel to a tertiary centre. Furthermore, compliance with California Children's Services Program standards (which include the availability of a paediatric cardiologist) is required for hospitals in California to be eligible for state reimbursement for NICU services. The availability of a paediatric cardiologist via telemedicine was the first time that a NICU has been permitted to maintain designation as a community level III nursery using telemedicine as its only means of staffing with a cardiologist. There were substantial financial benefits to the RH region of keeping live births requiring this level of care in local hospitals rather than transferring them out of the area for necessary level III NICU care.

Limitations

After the implementation of the telemedicine programme, a larger proportion of the infants admitted to the RH had echocardiograms (27% versus 40%). However, there was a noticeable increase in the annual rates (Table 1, Figure 1) in 2003, prior to the start of telemedicine. That is, other factors may have contributed to the increase. For example, an additional neonatologist was added to the RH staff and severity of illness also increased (data not shown). Therefore it is difficult to determine definitively the effect of telemedicine availability on the utilization of echocardiography. No increase in echocardiogram utilization was observed in previous studies.^6,19

There were other limitations to our data. First, due to the retrospective study design, evaluation of appropriateness of transfer was necessarily somewhat subjective. However identical criteria were applied to pre- and post-telemedicine groups. In addition, other factors may have influenced transfer patterns during the study period, including the establishment of paediatric cardiothoracic surgery programmes at two of the competing hospitals in the region (including the UCH) and expansion of echocardiography services at the UCH.

We did not collect data on time between study performance and interpretation in the pre- versus post-telemedicine periods, as the advantages of real-time and store-and-forward over mailed or couriered videotape have been well-documented.^3,6,9 Transfer times for the digital studies in this report were not recorded, but were as little as 15 minutes depending on size of the digital study (generally 35–100 MByte).

In conclusion, we found that the availability of telecardiology was associated with increased utilization of echocardiography, with an increase in the proportion of normal studies, and an increase in the percentage of neonates diagnosed with cardiac pathology without an increase in the number transferred for cardiac reasons. While health-care disparities will continue to exist between urban and rural or remote residents, our experience with telemedicine has shown it to be an effective way to bring subspecialty expertise to an isolated community, with benefits for both remote and tertiary centres and their patients.

Footnotes

Acknowledgements

Tannie Huang and Anita J Moon-Grady contributed equally to the paper. TH was supported by a Grant from the Children's Miracle Network.

References

Hoffman

, Kaplan

. The incidence of congenital heart disease. J Am Coll Cardiol 2002;39:1890–900

Finley

, Human

, Nanton

, Echocardiography by telephone – evaluation of pediatric heart disease at a distance. Am J Cardiol 1989;63:1475–7

Sobczyk

, Solinger

, Rees

, Elbl

. Transtelephonic echocardiography: successful use in a tertiary pediatric referral center. J Pediatr 1993;122:S84–8

Casey

, Brown

, Craig

, Rogers

, Mulholland

. Diagnosis of neonatal congenital heart defects by remote consultation using a low-cost telemedicine link. J Telemed Telecare 1996;2:165–9

Fisher

, Alboliras

, Berdusis

, Webb

. Rapid identification of congenital heart disease by transmission of echocardiograms. Am Heart J 1996;131:1225–7

Sable

, Roca

, Gold

, Gutierrez

, Gulotta

, Culpepper

. Live transmission of neonatal echocardiograms from underserved areas: accuracy, patient care, and cost. Telemed J 1999;5:339–47

Sable

. Digital echocardiography and telemedicine applications in pediatric cardiology. Pediatr Cardiol 2002;23:358–69

Sable

. Telemedicine applications in pediatric cardiology. Minerva Pediatr 2003;55:1–13

Sable

, Cummings

, Pearson

, Impact of telemedicine on the practice of pediatric cardiology in community hospitals. Pediatrics 2002;109:E3

10.

Awadallah

, Halaweish

, Kutayli

. Tele-echocardiography in neonates: utility and benefits in South Dakota primary care hospitals. S D Med 2006;59:97–100

11.

Finley

, Sharratt

, Nanton

, Paediatric echocardiography by telemedicine – nine years’ experience. J Telemed Telecare 1997;3:200–4

12.

Grant

, Wallace

, Hobson

, Craig

, Mulholland

, Casey

. Telemedicine applications for the regional paediatric cardiology service in Northern Ireland. J Telemed Telecare 2002;8 (Suppl. 2):31–3

13.

Justo

, Smith

, Williams

, Paediatric telecardiology services in Queensland: a review of three years’ experience. J Telemed Telecare 2004;10 (Suppl. 1):57–60

14.

Mulholland

, Casey

, Brown

, Application of a low cost telemedicine link to the diagnosis of neonatal congenital heart defects by remote consultation. Heart 1999;82:217–21

15.

Randolph

, Hagler

, Khandheria

, Remote telemedical interpretation of neonatal echocardiograms: impact on clinical management in a primary care setting. J Am Coll Cardiol 1999;34:241–5

16.

Tsilimigaki

, Maraka

, Tsekoura

, Eighteen months’ experience with remote diagnosis, management and education in congenital heart disease. J Telemed Telecare 2001;7:239–43

17.

Lewin

, Xu

, Jordan

, Accuracy of paediatric echocardiographic transmission via telemedicine. J Telemed Telecare 2006;12:416–21

18.

Rendina

, Downs

, Carasco

, Loonsk

, Bose

. Effect of telemedicine on health outcomes in 87 infants requiring neonatal intensive care. Telemed J 1998;4:345–51

19.

Scholz

, Kienzle

. Optimizing utilization of pediatric echocardiography and implications for telemedicine. Am J Cardiol 1999;83:1645–8