A telemedicine network for remote paediatric cardiology services in north-east Brazil

Abstract

Problem

Providing health care for children with congenital heart diseases remains a major challenge in low- and middle-income countries.

Approach

In October 2011, the Government of Paraíba, Brazil, established a paediatric cardiology network in partnership with the nongovernmental organization Círculo do Coração. A cardiology team supervised all network activities, using the Internet to keep in contact with remote health facilities. The network developed protocols for screening heart defects. Echocardiograms were performed by physicians under direct online supervision of a cardiologist; alternatively, a video recording of the examination was subsequently reviewed by a cardiologist. Cardiovascular surgeons came to a paediatric hospital in the state capital once a week to perform heart surgeries.

Local setting

Until 2011, the State of Paraíba had no structured programme to care for children with heart disease. This often resulted in missed or late diagnosis, with adverse health consequences for the children.

Relevant changes

From 2012 to 2014, 73 751 babies were screened for heart defects and 857 abnormalities were identified. Detection of congenital heart diseases increased from 4.09 to 11.62 per 1000 live births (P < 0.001). Over 6000 consultations and echocardiograms were supervised via the Internet. Time to diagnosis, transfers and hospital stays were greatly reduced. A total of 330 operations were carried out with 6.7% (22/330) mortality.

Lessons learnt

Access to an echocardiography machine with remote supervision by a cardiologist improves the detection of congenital heart disease by neonatologists; virtual outpatient clinics facilitate clinical management; the use of Internet technology with simple screening techniques allows resources to be allocated more efficiently.

Introduction

Caring for children with heart defects remains a challenge worldwide.¹ In developing countries, diagnoses are often late due to the lack of screening programmes and trained personnel.² The problem is worsened by limited availability of hospital beds and the remoteness of rural communities from main urban centres where paediatric cardiology specialists are available.³ Brazil faces all of these challenges, particularly in its poorest areas, the north and north-east parts of the country.⁴

Local setting

The State of Paraíba, located in north-east Brazil, has 3.7 million inhabitants. Around 70% of the children are cared for by the public health system; many live in rural areas and most come from very poor backgrounds. As there were no established paediatric cardiology facilities in Paraíba, children had to be referred outside the state for diagnosis and treatment. One of the main referral centres is located in the city of Recife, in the neighbouring state, Pernambuco (Fig. 1). Children were referred from towns and villages as far as 500 km from Recife; many arrived after a long time on a waiting list, with consequent deterioration of their clinical condition and some children died before being seen by the specialist.⁴

Fig. 1.

Health facilities in the Círculo do Coração paediatric cardiology network, Paraíba and Pernambuco, Brazil, 2014

Note: Recife and João Pessoa are the capital cities of the States of Pernambuco and Paraíba, respectively. Recife hosts the Círculo do Coração. João Pessoa hosts five network centres: the paediatric hospital, one level II and three level I maternity centres. Campina Grande is the largest city inland and hosts three centres: one level II and two level I. Patos hosts the third level II maternity unit. All other units are level I.

Approach

The need to improve this situation was evident and with scant existing resources and personnel, a novel solution had to be devised. Over the last two decades, telemedicine has proved to be an efficient tool for many point-of-care health applications.^5,6 In October 2011, a partnership programme was established between the Health Secretary of Paraíba and Círculo do Coração, a nongovernmental organization from Recife.

We conducted a review of routinely collected data⁷ on birth and death rates, socioeconomic conditions and the prevalence of heart defects in children living in Paraíba, from January 2001 to December 2011. Two initial priorities were defined: the establishment of a neonatal screening programme for the whole state and a hospital facility designated to manage patients.

Network structure

Initially, the 12 largest public maternity centres in the state were selected, together with one paediatric hospital. Centres were divided into three levels (designated I to III). All centres received tablet computers and pulse oximeters (level I); three maternity units also received a portable echocardiography machine (level II) and the paediatric hospital in the capital city of Paraíba State was equipped as a cardiology centre (level III). In 2014, further health centres were included in the network and training and consultation were expanded to include all aspects of perinatology (Fig. 1). A website was developed (https://www.circulodocoracao.com.br/sites/circor/en) and teleconference software was acquired. Three online clinics were established. Their purpose was to allow local paediatricians to examine children with heart defects with guidance from paediatric cardiologists via the Internet. These sessions aimed to reduce travel costs and provide a closer follow-up of children by the network.

A cardiology team was on-call 24 hours per day to supervise all network activities. The team consisted of 7 cardiologists, 3 residents and 4 staff (located in Recife). Three were specialized in paediatric echocardiography. The cardiology team performed daily rounds in all neonatal units from the participating sites, maintained intensive care unit supervision and organized teaching sessions, clinical and surgical meetings. A new perinatology team (with 13 neonatologists) joined the network in 2014. The perinatology team was mostly involved in teaching and seeing patients within the maternity centres. The health professionals were paid for the additional on call time – approximately 2000 United States dollars (US$) per month (exchange rate of 3 Brazilian Real to US$ 1) – by Círculo do Coração with funds from the Government of Paraíba.

Protocol development

Four initial protocols were developed by Círculo do Coração: (i) a training protocol, to explain the use of all equipment and software; (ii) a focused clinical examination protocol, to remind clinicians about the details of neonatal cardiology examination before discharge; (iii) a protocol for pulse oximetry testing of all babies born after 34-weeks gestation, based on guidelines published at the time⁸; and (iv) a screening echocardiogram protocol for neonatologists, which included three two-dimensional anatomical views and colour flow Doppler imaging.⁹ Members from all units were invited to participate in training sessions to learn and adhere to protocols. Each centre appointed three coordinators (one physician, one nurse and one computer support person) to report results and problems to the reference centre. The training protocol included an initial eight hour course followed by online sessions for all team members.

Screening tests

Indications for screening echocardiograms were either an abnormal clinical examination or pulse oximetry, defined as an oxygen saturation ≤ 95% or a difference in saturation greater than 2% between the right hand and one foot.¹⁰ Abnormal pulse oximetry results were automatically noted on a database, allowing the network to contact the clinic and request that they follow up any babies discharged home with abnormal test results. This active search protocol tracked over 80% (59 013/73 751) of the discharged neonates and ensured that abnormal findings were acted on.

Echocardiograms were done by neonatologists under direct online supervision by paediatric cardiologists, or a video recording of the examination was stored and forwarded together with the neonatologist’s initial diagnostic impression. Paediatric cardiologists reviewed and reported on the videos, with advice on clinical management, within one day. Virtual outpatient sessions, ward rounds and other meetings were also scheduled to provide a full range of interactions between the health workers in rural areas and smaller municipalities in Paraíba and the paediatric cardiologists at the reference centre.

Surgeons and anaesthetists from Recife agreed to travel to the paediatric hospital in João Pessoa, the capital city of Paraíba, once a week, to perform heart surgery. The more complex cases, however, were referred to Recife.

Technical specifications

Internet connections were unreliable for some health centres. To overcome this problem, tablet computers with third generation mobile wireless Internet connections were distributed to all centres. Webex teleconference software (WebEx Communications Inc., Milpitas, California) was acquired to provide secure communication over the Internet. Online meetings were held each day, among all centres, using existing tablets or laptop computers. Echocardiogram images were either directly acquired from the echocardiogram screens or stored and subsequently uploaded to the website.

Relevant changes

In total, 76 374 patients were seen from January 2012 to December 2014. This included 190 pregnant women (0.2%); 73 751 neonates (96.6%) and 2433 older children (3.2%) with a mean age of 3.04 ± 3.77 years, (range: 30 days to 17.5 years). This represents approximately 80% (73 751/91 615) of the target population (neonates with 34 or more weeks of gestational age in the participant centres) and over 60% (73 751/120 484) of all births in the public health system in the state. There were 1320 abnormal pulse oximetry tests and 1067 abnormal findings on clinical examination of the cardiovascular system; in 77 cases, both pulse oximetry and clinical examination were abnormal.

Initially, all echocardiograms were done with online supervision by the paediatric cardiologist, as part of the neonatologists’ training. After performing about 100 examinations, the quality of images obtained became significantly better and the operators were more confident. At this point, the cardiologists waited for requests for direct online supervision, which dropped progressively until being sought only when pathological findings were suspected. As there are always new neonatologists being trained, this learning process and interaction between teams is a continuous cycle.

There were 1815 screening echocardiogram tests done, of which 848 were abnormal, 957 were normal and 10 were inconclusive. However, 2310 children had indications for a screening echocardiogram. The difference, 495, was mainly due to false-positive oximetry results early in the development of the network. If an echocardiogram was inconclusive, the diagnosis was subsequently established by echocardiography done by a paediatric cardiologist. From the abnormal and inconclusive echocardiograms, 857 demonstrated congenital heart disease (11.62 per 1000 live births). Neither a patent foramen ovale nor an isolated, small, arterial duct was considered a congenital heart defect. However, a clinically significant patent ductus arteriosus was included, coded as transitional circulation. The prevalence of eight major congenital heart defects before and after the introduction of the cardiology network is compared with previously published data in Table 1.^11,12

Table 1. Birth prevalence of the most common subtypes of congential heart disease for major country groups (1970–2010) and for Paraíba, Brazil (2001–2011 and 2012–2014).

Type of defect	Prevalence of defect per 1000 births								Pc
	Africa	Asia	Europe	North America	Oceania	South America	Paraíba, Brazil
							2001–2011a	2012–2014b
Atrial septal defect	0.35	1.71	1.66	1.71	0.47	0.70	0.17	1.19	< 0.01
Ventricular septal defect	1.40	2.47	2.71	2.42	2.56	1.86	0.71	3.62	< 0.01
Patent ductus arteriosus	0.45	0.67	0.94	0.50	0.45	0.40	1.10	4.53	< 0.01
Pulmonary stenosis	0.28	0.68	0.50	0.41	0.40	0.36	0.27	0.26	0.96
Tetralogy of fallot	NR	0.42	0.33	0.34	0.31	0.37	0.18	0.21	0.83
Coarctation of the aorta	0.06	0.20	0.34	0.30	0.60	0.30	NR	0.17	< 0.01
Transposition of great arteries	0.67	0.18	0.34	0.25	0.38	0.19	0.13	0.21	0.34
Aortic stenosis	NR	0.08	0.25	0.18	0.18	0.08	0.03	0.04	0.90

^a Period before implementation of the paediatric cardiology network.

^b Period after implementation of the paediatric cardiology network.

^c  P-value of the comparison between Paraíba 2001–2011 and Paraíba 2012–2014.

Data sources: Christianson et al.¹¹ and van der Linde et al.¹²

In the five online clinics supervised by the cardiologists in the network, 1092 patients had over 6000 consultations and echocardiograms. A total of 330 operations were done; 285 in João Pessoa and 45 in the referral centre in Recife. There were 30 neonates (9.1%), 65 infants (19.7%), 78 toddlers (23.6%) and 157 older children (47.6%). The overall mortality of 6.7% (22/330) was within the expected range for developing programmes. Mortality risk was related to surgical complexity and clinical condition according to Rach’s score¹³ and a post-operative index.¹⁴ Time between birth and diagnosis was less than three days in most cases, with a maximum of 647 days (due to the late clinical presentation of milder forms of congenital heart disease). Hospital transfers as well as hospitalization periods were reduced as children did not have to wait to be transferred for echocardiagrams and operations. The virtual clinics were used to facilitate local follow-up for most patients. There were no cases of medical litigation involving the management of children with congenital heart disease.

The total cost for establishing and operating the network was US$ 1.2 million in the first year. With the expansion to a total of 21 centres and perinatology services in 2014, the annual cost increased to US$ 2.0 million. A more detailed study of the economic impact, including the impact of perinatology services, is being conducted. The initial impact of cardiology services was estimated in comparison with the number of patient transfers outside the north-east area, detection rates for congenital heart defects and litigation costs (details are available from the corresponding author).

Lessons learnt

Several problems were encountered during development of the network, including inadequate equipment, overloaded clinical settings and local changes in nursing staff with insufficient training of new members of staff. The wide range of health workers using the new technologies was another problem. Local training on the equipment was therefore done on a regular basis in addition to the online training. Access to an echocardiography machine by neonatologists with direct online supervision was the most important factor leading to improved diagnosis of congenital heart disease (Box 1). This screening model is similar to others,^15,16 but its impact was probably greater, due to the previous lack of paediatric cardiologists in this population.

Box 1. Summary of main lessons learnt.

• Access to echocardiograph facilities with online supervision improves the detection of congenital heart disease in this rural setting.

• Online outpatient clinics facilitate clinical management.

• The combination of simple screening techniques and diagnostic technology allows resources to be allocated more efficiently.

Clinical care for the children was a big challenge. Online outpatient clinics were a major facilitator of clinical management, by reducing the need for transportation, empowering local physicians and involving other professionals in patient care. However, children requiring surgery had to enter waiting lists to be directed either to the paediatric hospital in the state capital or to Recife. In conclusion, through both live and online collaborative work, local professionals were able to screen, diagnose and treat children with congenital heart disease from remote areas.