Peer-to-peer tele-consultative services for critical care, Afghanistan, Kenya, Pakistan, United Republic of Tanzania

Abstract

Objective

To develop a tele-intensive care service providing peer-to-peer teleconsultation for physicians in remote and resource-constrained health-care settings for treatment of critically ill patients, and to evaluate the outcomes of the service.

Methods

The Aga Khan University started the coronavirus disease 2019 (COVID-19) tele-intensive care unit in 2020. A central command centre used two-way audiovisual technology to connect experienced intensive care specialists to clinical teams in remote hospital settings. The service, always available, used messaging applications and telephone calls. Coverage was later extended to other medical, neonatal, paediatric and surgical patients requiring critical care.

Findings

Between June 2020 and December 2023, the service provided 6014 teleconsultations to manage 1907 patients in 109 medical facilities, mostly in Pakistan and also Afghanistan, Kenya and United Republic of Tanzania. Of the 1907 patients, 652 (34.4%) had COVID-19 and 1244 (65.6%) had other illnesses. The mean duration of teleconsultations was 14.5 min. Of 581 patients for whom outcome data were available, 204 (35.1%) died. Multivariate multinomial logistic regression showed the odds of death decreased with increased number of consultations (> 3) per patient (adjusted odds ratio (aOR): 0.28; 95% confidence interval, CI: 0.16–0.48), and increased number of recommendations (≥ 5) per consultation (aOR: 3.09; 95% CI: 1.08–8.84).

Conclusion

Our tele-intensive care service helped manage critically ill patients in regions where intensive care had not previously been available. While research on the clinical impact of this model is needed, decision-makers should consider its use to increase provision of critical care in remote and resource-constrained health-care settings.

Introduction

The effective functioning of intensive care units requires staffing with nurses, intensive care physicians and allied health professionals. Research suggests that adequate staffing is associated with reduced intensive care unit mortality and shorter length of stay.^1–5 However, intensive care unit capacity, including beds, equipment and staff, can be limited and even non-existent in low- and middle-income countries. As of 2020, low-income countries had between 0.0 and 21.3 intensive care unit beds per 100 000 population, compared with 0.0 to 59.5 per 100 000 population in high-income countries.³ Globally, at least 96 countries and territories had fewer than 5.0 intensive care unit beds per 100 000 population.³ Most intensive care units in these places are in large urban referral hospitals, making it challenging for patients in rural areas to access critical care services.

In Pakistan, the reported critical care capacity is low for a population of about 227 million in 2020.^6,7 A recent study found that only 36.8% (39/106) of hospitals in Pakistan employed accredited intensive care physicians.⁸ As in other low- and middle-income countries, most of Pakistan’s population (62.0%) lives in remote and rural locations and has limited access to appropriate medical care.⁹ This lack of access is because of the greater difficulty in providing care outside urban areas and the reluctance of physicians to work in these settings.^8–10

Technological innovation has been used to meet the demand for intensive care unit services in low- and middle-income countries and overcome the shortage of intensive care specialists in many remote settings.¹¹ Intensive care telemedicine (tele-intensive care) is a technology-based model designed to connect remotely located physicians with expert intensive care clinicians via technology such as video conferencing.^12–14 Because tele-intensive care helps bedside physicians diagnose and treat the most acute hospital patients, this service has been associated with substantial reductions in mortality, shorter stays in hospital and the intensive care unit, and lower health-care costs.^12–14 Traditional approaches to tele-intensive care are costly and require an advanced technology infrastructure which is usually proprietary, such as remotely accessible electronic medical records and live telemetry feeds for physiological data. Therefore, these systems are used more often in high-income countries¹⁵ than in low- and middle-income countries.¹⁶

Gaps in critical care capacity in low- and middle-income countries adversely affected the ability of health services to cope with surges in critically ill patients with coronavirus disease 2019 (COVID-19). Therefore, the Aga Khan University developed and implemented a tele-intensive care unit to address the lack of skilled personnel in resource-constrained health-care settings. The aim of this intervention was to bridge the gap between well-resourced and less-resourced intensive care facilities during the COVID-19 pandemic and beyond. The tele-intensive care unit uses two-way audiovisual communication through available devices, such as smartphones, coupled with computer systems to provide a collaborative, interprofessional care model focused on managing critically ill patients.^7,17–19 In this study, we outline the processes involved in establishing tele-intensive care, and assess the effect of tele-intensive care on patient care and outcomes.

Methods

In June 2020, the Aga Khan University started the COVID-19 tele-intensive care unit project in collaboration with the federal governments of several countries and provincial governments in Pakistan. The objective of the tele-intensive care unit was to increase access of physicians in remote areas to intensive care experts who provided evidence-based guidance with the aim of improving patient outcomes by standardizing processes.^20–22 This unit offered round-the-clock expert consultations for critically ill patients in Afghanistan, Kenya, Pakistan and United Republic of Tanzania. The Pakistani health ministry approved the project. The Aga Khan University Ethical Review Board approved the study (2020–4869–10701).

Development of the tele-intensive care unit

A project team consisting of a principal investigator, two co-principal investigators, a medical lead, a project manager, an assistant manager, and medical and research staff created a tele-intensive care central command centre to provide access to peer-to-peer consultation.

Governance and management

The project team engaged with critical care stakeholders in public and private health-care sectors through the federal and provincial governments and health ministries across Pakistan to create the tele-intensive care hub. The federal and provincial governing bodies and the Aga Khan University signed memoranda of understanding outlining the scope, nature, limitations and expectations of the project, which enabled communication with public sector medical institutions across Pakistan. The government entities disseminated official notices about the intervention to applicable medical facilities to encourage their participation. Simultaneously, the project team ran a social media campaign announcing the tele-intensive care intervention targeting private sector institutions. All institutions willing to participate signed an agreement outlining roles, responsibilities, scope and liabilities. Each institution nominated a senior clinician as the focal person for coordination and stakeholder management. This person liaised with the tele-intensive care unit to identify local physicians working in intensive care who might ask for the consultations, facilitate implementation of consultations and overcome day-to-day operational challenges. Institutions outside Pakistan were approached about participating in the second year of the project.

Staffing and training

The Aga Khan University assembled a multidisciplinary team of up to 16 tele-intensive care consultants. The team included anaesthesiologists, intensive care physicians, internal medicine physicians, neonatologists and paediatricians. The project team developed monthly workflow schedules and rotations, and consultants were paid a small stipend for each consultation. Additionally, the project leads recruited recent medical graduates who were trained as tele-intensive care officers to facilitate communication between the remote physicians and the tele-intensive care consultants. Details of their training are published elsewhere.²³ These officers continuously monitored the telephones and WhatsApp (Meta, Menlo Park, United States of America) groups for consultation requests, documented the patient care recommended from the consultations and maintained the tele-intensive care patient registry. A manager oversaw stakeholder management, external and internal communication, data management and the operational processes.

Tele-intensive care unit framework

The tele-intensive care officers were located either inside or outside the command centre. They connected the tele-intensive care consultants to medical facilities requesting consultations, or arranged remote consultation through computers or mobile devices with videoconferencing capability from any location with internet access.

Consultative services could be provided at predetermined times at the convenience of the requesting physicians and tele-intensive care consultant (scheduled-care model). Alternatively, teleconsultations were prompted by an alert through which the requesting physician shared the patient’s condition with the tele-intensive care officer to get an immediate response (responsive-care model). A hotline number was assigned for such calls.

Teleconsultation process

Remote physicians initiated teleconsultation requests through text messages or telephone calls (Fig. 1). Thereafter, WhatsApp, telephones, mobile telephones and Zoom (Zoom Communications Inc., San Jose, USA) platforms were used for communication as they were readily available to the remote physicians. Relevant patient medical data, such as radiographs, laboratory tests, medications, ventilation management and acute adverse events in the past 24 hours, were sent from the remote physician to the tele-intensive care officer. The officer then arranged a three-way teleconsultation between themself, the requesting physician and the tele-intensive care consultant.

Fig. 1.

Steps in the tele-intensive care unit consultation process

Patients

The patients referred for tele-intensive care were inpatients in the high-dependency or intensive care unit of the remote facilities whom the bedside physician thought needed critical care consultation. Initially, consultations were provided for COVID-19 patients; however, after 12 months, coverage was broadened to include other surgical, medical, neonatal and paediatric critical care patients.

Promotion and dissemination

Social media platforms, the Aga Khan University’s official webpage and e-flyers distributed via emails (to people who had previously registered for any education course offered by the university) were used to raise awareness of the tele-intensive care service and hotline. We also initially used site visits to medical facilities across Pakistan to promote the service and build relationships.

Data management

The tele-intensive care officer provided a summary of the consultation, including patient assessment and a management plan recommended by tele-intensive care consultants to the requesting remote physician. Details of the teleconsultation (i.e. patient demographics, admission symptoms, investigations and medications discussed, management recommendations, and consultation times and method) were documented in the tele-intensive care patient registry. Investigations discussed were categorized into haematology, biochemistry, radiology and COVID-19-related tests. Management recommendations were categorized into diagnostic interventions (e.g. recommendations for further testing) and therapeutic interventions (e.g. pharmaceutical and critical care supportive management). Only tele-intensive care officers and their supervisors had permission to access and edit these data. Patients were assigned a unique medical record number and data were de-identified. The tele-intensive care consultants addressed any subsequent queries through text-based communications, which were recorded by tele-intensive care officers. The officers conducted follow-ups after 24–48 hours, with further follow-up teleconsultations scheduled if required.

Funding

The project was funded by in-kind funding by the Aga Khan University and independent grants from agencies listed in the funding section. The university support included space for the hub and faculty time. The agency funding covered all programmatic costs including equipment, project personnel, required travel and meetings and consultant stipends, and educational content development and delivery. The consultation service was free for the remote institutions.

Data analysis

We used an observational quantitative design to evaluate the care provision processes and assess outcomes of patients. The project team retrieved data for the number of new COVID-19 patients in Pakistan from the Government of Pakistan’s official website. We present data as numbers and percentages for categorical variables. We did a stratification analysis of intensive care unit outcomes. We used multivariate multinomial logistic regression analysis to assess factors associated with intensive care unit outcome. P < 0.05 was considered statistically significant. The reference variable was discharged home. We used Excel, version 2005 (Microsoft, Redmond, USA) and R 4.1.2 (R Foundation, Vienna, Austria) for data analysis. For the costing analysis, we used actual incurred costs for programmatic start-up, and fixed and variable operational costs based on United States dollars (US$) of April 2023.

Results

The tele-intensive care unit provided 6014 consultations for 1907 patients from 1 June 2020 to 31 December 2023, with an average of 139.9 consultations per month. A total of 1138 (59.7%) patients received one consultation and 769 (40.3%) received more than one. Most patients referred for tele-intensive care were male (1016: 53.6%) and lived in Pakistan (1858: 98.3%). Of the 5291 consultations conducted with known type of consultation, 2728 (51.9%) were scheduled and 2533 (48.1%) were responsive. Of the 652 (34.4%) patients presenting with COVID-19, 344 had severe symptoms and 172 had critical symptoms;^24,25 1244 (65.6%) did not have COVID-19 (Table 1). The highest number of teleconsultations was in August 2021 (Fig. 2; online repository).²⁶ The most common active medical issue for which consultations were requested were respiratory conditions (5515/8436 requests; 65.4%) and neurological disorders (634/8436 requests; 7.5%; Table 2).

Table 1. Characteristics of consultations and patients referred to the tele-intensive care unit, June 2020 to December 2023.

Variable	No. (%)a
Total no. of consultations	6014
Type of consultation
Scheduled	2728 (51.9)
Responsive	2533 (48.1)
Unknown	753 (12.5)
Total no. of patients	1907
No. of consultations per patient
1	1138 (59.7)
> 1	769 (40.3)
Sex of patient
Male	1016 (53.6)
Female	878 (46.4)
Unknown	13 (0.7)
Age, in years
< 18	397 (21.6)
18–38	476 (25.9)
39–58	395 (21.5)
59–78	458 (24.9)
79–98	112 (6.1)
> 98	2 (0.1)
Unknown	67 (3.5)
Location of health facility
Afghanistan	9 (0.5)
Kenya	8 (0.4)
Pakistan	1858 (98.3)
Baluchistan	5 (0.3)
Gilgit Baltistan	353 (19.0)
Khyber Pakhtunkhwa	351 (18.9)
Chitral	88 (4.7)
Punjab	713 (38.4)
Sindh	348 (18.7)
United Republic of Tanzania	16 (0.8)
Unknown	16 (0.8)
Diagnosis on admission
COVID-19	652 (34.4)
Asymptomatic	37 (5.7)
Non-severe	76 (11.7)
Severe	344 (52.8)
Critical	172 (26.4)
Suspected	23 (3.5)
Other	1244 (65.6)
Unknown	11 (0.6)

^a Percentages are calculated from the total sample minus the number of unknowns. The unknown percentages are calculated from the total sample.

Fig. 2.

Number of patients and consultations and new COVID-19 cases reported in Pakistan, by month and year, June 2020 to December 2023

COVID-19: coronavirus disease 2019.

Table 2. Types of medical problem for which remote physicians requested tele-intensive care consultations, Afghanistan, Kenya, Pakistan, United Republic of Tanzania, June 2020 to December 2023.

Type of medical problema	No. (%) (n = 8436)
Cardiological	454 (5.4)
Dermatological	96 (1.14)
Endocrine	176 (2.1)
Ear, nose and throat	15 (0.2)
Gastrointestinal	286 (3.4)
Gynaecological	6 (0.1)
Haematological	144 (1.7)
Infectious diseases	387 (4.6)
Metabolic disorders	291 (3.4)
Musculoskeletal	47 (0.6)
Renal	259 (3.1)
Neurological	634 (7.5)
Psychological	59 (0.7)
Respiratory	5515 (65.4)
Other	67 (0.8)

^a A patient could have had more than one medical problem.

Teleconsultations were provided to 109 medical institutions, 105 in Pakistan, three in East Africa (two in Kenya, one in United Republic of Tanzania) and one in Afghanistan. The average number of consultations per patient was 3.2 in Pakistan, 2.4 in East Africa and 2.1 in Afghanistan. The institutions included three primary care facilities, 47 secondary care facilities, 51 tertiary care facilities and one temporary COVID-19 testing and vaccination facility; for seven facilities the type was unknown. Of the 109 facilities, 29 were in rural areas, 77 were in urban areas, with the location of three facilities unknown.²⁷

The total consultation call duration was 77 981 minutes, mean duration of 14.53 minutes/call (online repository).²⁶ For the 6014 consultations, voice calls were the most frequently used consultation method (4368; 72.6%; online repository).²⁶

The tele-intensive care consultants reviewed 16 166 investigations (laboratory tests and radiological images), with a mean of 8.5 investigations reviewed per patient and 2.7 investigations per consultation. The investigations were related to biochemistry (9310; 57.6%), haematology (3499; 21.6%), radiology (2233; 13.8%), COVID-19 investigations (591; 3.7%), microbiology (484; 3.0%) and other investigations (49; 0.3%). The most frequent specific investigations were complete blood count (2734), creatinine level (2013) and arterial blood gases (1971; online repository).²⁶

Tele-intensive care consultants made 37 904 management recommendations, with a mean of 19.9 recommendations per patient and 6.3 recommendations per consultation. The most management recommendations (3897) were provided during August 2021. The most frequently recommended management advice related to therapeutic techniques to manage infectious diseases and infection control (6752 recommendations; online repository).²⁶

Outcome data were available for 581 (30.5%) patients, 204 (35.1%) of whom died (online repository).²⁶ Multivariate multinomial logistic regression analysis showed no significant relationship between patient outcome and sex. The odds of death decreased significantly with increased number of consultations (> 3 consultations) per patient (adjusted odds ratio, aOR: 0.28; 95% confidence interval, CI: 0.16–0.48). Patients who received more than five recommendations had significantly greater odds of death than patients with fewer recommendations (aOR: 3.09; 95% CI: 1.08–8.84; Table 3).

Table 3. Factors associated with intensive care unit outcome.

Variable	aOR (95% CI)
	Discharged from intensive care unit	Died	Left health-care facility against medical advice	Transferred
Sex
Female	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)
Male	0.98 (0.58–1.70)	1.26 (0.80–1.98)	0.77 (0.27–2.18)	1.04 (0.57–1.92)
No. of consultations
1	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)
2–3	0.57 (0.27–1.20)	0.76 (0.40–1.48)	0.88 (0.19–4.20)	0.46 (0.20–1.10)
> 3	0.20 (0.10–0.37)	0.28 (0.16–0.48)	0.43 (0.12–1.57)	0.23 (0.11–0.46)
No. of laboratory tests
0	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)
1	1.16 (0.32–4.31)	0.91 (0.36–2.32)	0.28 (0.03–3.12)	0.66 (0.18–2.37)
2–4	2.02 (0.65–6.25)	1.29 (0.58–2.84)	0.61 (0.13–2.88)	1.13 (0.40–3.22)
≥ 5	2.97 (0.94–9.27)	1.38 (0.62–3.12)	0.44 (0.09–2.24)	0.95 (0.33–2.81)
Mean time of consultation (minutes)
< 15	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)
15–30	0.96 (0.51–1.78)	1.15 (0.69–1.91)	1.30 (0.41–4.07)	0.89 (0.43–1.83)
> 30	0.15 (0.02–1.34)	0.85 (0.32–2.30)	0.70 (0.07–7.08)	0.66 (0.15–2.84)
No. of recommendations
0	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)
1	3.23 (0.51–20.00)	1.71 (0.38–7.59)	4.35 × 10−9 (8.25 × 10−10–8.30 × 10−9)	1.55 (0.18–13.15)
2–4	1.63 (0.38–6.99)	2.50 (0.86–7.32)	0.22 (0.02–2.96)	1.63 (0.38–7.18)
≥ 5	1.43 (0.34–6.01)	3.09 (1.08–8.84)	1.49 (0.23–9.49)	1.65 (0.38–7.10)
Diagnosis on admission
Asymptomatic COVID-19	1.00 (reference)	1.00 (reference)	1.00 (reference)	1.00 (reference)
Critical COVID-19	0.27 (0.01–4.75)	2.55 (0.30–21.50)	48 057.44 (17 950.98–128 656.90)	0.36 (0.02–5.76)
Severe COVID-19	0.97 (0.07–12.79)	0.94 (0.12–7.62)	8 951.41 (3 547.05–22 589.99)	0.55 (0.04–7.02)
Non-severe COVID-19	0.20 (0.01–5.70)	0.13 (0.01–2.47)	1.0 × 10−5 (1.4 × 10−6–6.28 × 10−5)	0.19 (0.01–5.23)
Non-COVID-19 illness	2.28 (0.19–30.73)	1.96 (0.24–16.11)	23 515.98 (9 120.56–60 632.37)	1.58 (0.12–20.24)

CI: confidence interval; COVID-19: coronavirus disease 2019; aOR: adjusted odds ratio.

Note: We used a multivariate multinomial logistic regression to calculate aOR. Discharged home was the reference variable.

Initial capital costs of the project, including computers, mobile telephones, accessories such as webcams and headsets, and a Zoom account were US$ 9467. Ongoing fixed operational costs, including personnel salaries (physician project lead, management and administrative staff, and tele-intensive care officers) and consumables, were aggregated for a typical month and came to US$ 168.88 per day. Ongoing variable operational costs, mainly for consultant stipends, were US$ 18.20 per patient.

Discussion

Telemedicine can facilitate patient stabilization by providing immediate virtual assistance while using the knowledge and skills of the bedside physicians.^12,20 Peer-to-peer telemedicine has shown success, particularly in high-income countries, with reduced intensive care and hospital stay and mortality reported.^20,28,29 Research also shows that a lack of critical care staff to meet needs can be bridged by providing teleconsultation services.^30–32 However, previous efforts have required extensive and often costly technological resources to deliver these services.^7,33–35 Our results show that a tele-intensive care unit was able to reach remote intensive care facilities using existing telecommunication and information technology. They also demonstrate the feasibility of using this model in low- and middle-income countries to provide consultative services to facilities with limited resources, trained critical care staff and infrastructure.

During the COVID-19 pandemic, the number of teleconsultations mirrored the national COVID-19 numbers and waves. After the pandemic, there was an initial lull in teleconsultations but these began to rise again in mid-2023. These consultations were likely for other kinds of critically ill patients and as a result of wider recognition of this service. Bedside physicians most frequently requested teleconsultations for management strategies for respiratory complications. Severe and critical diseases can lead to respiratory failure and acute respiratory distress syndrome, which require rapid implementation of interventions such as intubation and mechanical ventilation that many physicians are not comfortable using.³⁶ This observation aligns with the fact that 79.1% (516/652) of the patients with COVID-19 presented with severe or critical symptoms which generally require advanced respiratory support. Severe respiratory problems remained the leading reason for consultation after the pandemic ended.

Tele-intensive care has reduced intensive care and hospital mortality, further endorsing the validity of these services.^{16,20–22,28,29} In the United States, tele-intensive care services reduced the hospital mortality from 13.6% (208/1529) before the intervention to 11.8% (562/4761) after implementation.¹ Our service recorded a 35.1% (204/581) mortality (among patients for whom outcome data were available); and our multivariate multinomial logistic regression analyses indicated that the more teleconsultations provided per patient, the greater the likelihood of patient survival. This mortality was similar to that in critically ill patients in a tertiary care facility in metropolitan Pakistan in a 90-day study (39.4%; 67/170).³⁷ This finding is noteworthy given the differences between a teleconsultation model and in-person model for patient care, as well as the differences in resources between the two models. A teleconsultation model can also be more easily scaled up than one based on in-person care. While further study about causality and clinical impact of this model is needed, decision-makers should consider its use to increase provision of critical care in resource-constrained settings.

Our service offered peer-to-peer patient management supported 109 facilities in Afghanistan, Kenya, Pakistan and the United Republic of Tanzania. A few primary care facilities, which usually do not have intensive care units, used our services when COVID-19 started, as they were the first health-system access point for ill patients. To broaden the reach and impact of such a teleconsultation service, primary care facilities should be approached for collaboration as they often provide initial acute care and referrals for critically ill patients. Measures might be needed to provide the necessary technology, as many do not have access to high-speed internet and may not have the technology for virtual consultations.

A limitation of our study was the potential for selection bias towards minimizing the impact on patient outcomes, given that bedside physicians might have used this tele-intensive care service for their sickest patients, potentially leading to a higher mortality rate. Additionally, the main objective of the service was provision of tele-consultations for critically ill patients to support staff during the pandemic and beyond; as such, patient outcomes were not the focus of the project. Therefore, not all patients could be followed up, and less than one third of the patients for whom consultations were provided were included in our analysis. Since most of the patients lost to follow-up had only one consultation, the effect on the outcomes of critically ill patients would be limited.

The tele-intensive care service faced various implementation and sustainability challenges. We conducted a preliminary study (unpublished data) of bedside physicians’ perception of tele-intensive care. Despite offering free peer-to-peer consultations and receiving positive feedback during COVID-19 surges, some physicians, particularly senior doctors, perceived them as a threat to their practices and autonomy, and did not allow junior doctors in their units to start consultations. Furthermore, these physicians were reluctant to disseminate information about the service in their units. Additionally, most patients had only one consultation. Another challenge was obtaining patient outcomes after the patient became stable or was discharged from the intensive care unit. Because tele-intensive care services were provided free-of-charge for patients and bedside physicians received no financial support for teleconsultation, these physicians might have seen daily consultation as an additional burden. Consequently, they may have been reluctant to provide extensive documentation to the Aga Khan University and shared only limited information.

To address these challenges, we held meetings with hospital staff and maintained constant engagement through educational activities and e-mentoring for bedside physicians. These efforts resulted in sustained engagement with several institutes. Indeed, 31 of the 109 institutes are still using our tele-intensive care services. Our medical officers also conduct daily follow-up calls after teleconsultations to assess patient outcomes and promote further consultations for each patient. Additionally, they document all patient health information during audioconferencing to reduce the documentation burden on the bedside physicians.

The COVID-19 pandemic increased the use of telemedicine to deliver care to remote and resource-constrained facilities. Our tele-intensive care service has successfully helped manage critically ill patients since June 2020 in regions where tele-intensive care had not previously been used. This project has continued due to its low implementation cost per patient, continuous availability of intensive care doctors, and use of existing telecommunication and information technology infrastructure. Future research and coordination are necessary to increase access to the internet and technology required for teleconsultations and to determine the effect of the project on patient mortality.