To accomplish our first research objective, we performed POCT needs assessment in Cambodia on the basis of prior field research in limited-resource settings [19-24] and methods codified by Point-of-Care Technologies Centers and Research Network investigators and funded by the US National Institute of Biomedical Imaging and Bioengineering at the National Institutes of Health [25-27]. Considering the widespread reach of COVID-19, we surveyed academicians, professionals, business operators and employees, and community laypersons to assess the potential for use of rapid antigen tests (RagTs), self-testing and mobile diagnostics, and the suitability of travel to reverse transcriptase polymerase chain reaction (RT-PCR)–testing referral sites.

We generated a questionnaire in English and the Khmer language (available in Multimedia Appendix 1) to rapidly gather data related to the challenging conditions of episodic quarantines, red zones of high risk, and prolonged lockdowns. We used the structured questionnaire to systematically and rapidly identify and collate needs through meetings, university exchanges, webinar discussions, lecture follow-up and feedback, Zoom sessions, facility inspections, field interviews, and a focus group (13 participants). Overall, 137 participants contributed to the assessment survey and compilation of needs.

Contributing survey respondents comprised (1) public health, clinical laboratory, field testing, and medical personnel including professors, physicians, nurses, pharmacists, a public health analyst, medical technologists, a quality control supervisor, emergency medical system operators, and private laboratory staff; (2) a civil servant, project manager, and biologist; (3) professional leaders comprising the President of the Cambodian National Association of Medical Technologists, the faculty of the University of Puthisastra, and the director, workers, and students from the Ministry of Health National Public Health Laboratory and the National Institute of Public Health in Phnom Penh; (4) road checkpoint officials in provinces, tourist agents, community workers, transportation staff, hotel personnel, Angkor Wat employees in Siem Reap Province, and personnel working in nongovernment organizations; and (5) the Phnom Penh City pandemic committee that monitored outbreaks and community police officers who enforced quarantines.

We augmented respondent data by making numerous telephone calls; by engaging manufacturers of tests (in Phnom Penh, Singapore, South Korea, and Thailand); by monitoring commercial sales status in Cambodia; through email correspondence; and by merging the data provided by local health care experts in Banteay Meanchey, other border provinces, and Phnom Penh. The residents and family members in Siem Reap Province—a site of high rural infection rates that shares heavily used transportation routes with the Banteay Meanchey border province to the west—were queried as well as Cambodian workers, some of whom were stranded when Thailand, under duress from increasing contagion, unilaterally closed border gateways.

To accomplish our second research objective, we used geospatial science techniques innovated for limited-resource settings by Ferguson et al [28-30] and Kost [31,32]. This was done to metricize distances from the main villages of each district within the Cambodian border provinces to COVID-19 testing sites at province referral hospitals. Transit times were calculated by dividing the distances traveled by the speed limit of 30 km/h for motorcycles in towns. In general, rural residents did not have access to automobiles or buses for routine transport.

Figure 1 identifies provinces, international borders, bilateral crossing gates, border closures, and lockdowns. Physical locations were based on data obtained from web-based open-source mapping software (Cambodia Covid Restrictions Map 2021 MapItKH [33], OnTheWorldMap [34]), hardcopy Cambodia and Thailand maps, web resources [National Institute of Public Health, Phnom Penh Post, Khmer Times, digital libraries], and literature searches (PubMed, Google). Population census, national health care data, and public health reports (Open Development Cambodia [35]) helped determine public accessibility to local COVID-19 testing sites.

Province data are reported as median distance and median time to testing for each province in descending-value Pareto plots, which also identify 20% thresholds for cumulative distance. Differences were assessed using the Kruskal-Wallis test followed by pairwise analysis (using the Mann-Whitney U test). Paired differences (maxima minus minima) were analyzed using the Wilcoxon signed rank test. A value of P<.05 was considered significant. Data were not distributed normally (Shapiro-Wilk test). However, for Banteay Meanchey Province, we reported the median (range) and mean (SD) distances to testing and the transit time to document distribution skewness.

Popular COVID-19 RAgTs in Southeast Asia at the time of the peak outbreaks in Cambodia included the Roche Diagnostics SARS-CoV-2 RAgT and the Abbott Panbio RAgT. Mathematical analyses [10,36-39] were conducted in parallel with field research in Cambodia to (1) determine if these two tests met the World Health Organization (WHO) criterion for sensitivity >80% (median) in clinical evaluations, (2) characterize performance patterns and their dependence on prevalence, and (3) support strategic recommendations in the discussion. Tables S1 and S2 in Multimedia Appendix 1 tabulate the clinical evaluations, Figures S1 and S2 in Multimedia Appendix 1 compare performance plots, and Table S3 in Multimedia Appendix 1 lists Bayesian formulas derived to graph performance.

To improve accessibility to diagnostics, a COVID-19 RAgT vending machine was placed outside the library in Esparto, the first site in California to implement automated, 24-7, free public access to testing, which has expanded now to include free masks, tick and insect repellent, male and female condoms, and naloxone. Esparto is a small rural community of 4009 people near the western foothills of Yolo County. Additional county sites were established outside the Davis Library, West Sacramento Community Center, Winters City Hall, and La Superior Market in Woodland.

One could obtain as many test kits (eg, Flowflex; ACON LABS; Bio-Self, Bioteke USA) as needed by simply tapping a key on the vending machine. There was no need for identification, registration, or follow-up. Asymptomatic individuals could also obtain free RAgT test kits (InteliSwab; Orasure Technologies) from staff inside county public libraries. Local access continues through 2024; the US government has ended free RAgT access online, having distributed over 900 million test kits, but still provides them to uninsured individuals and underserved communities. Papers by Kost [37,40] illustrate storyboards of the self-testing process steps. The National Institutes of Biomedical Imaging and Bioengineering (NIH Radx Tech Initiative) implemented a “Make My Test Count” website for anonymous reporting of self-testing results [41].

The Ethics Committee at the University of Puthisastra, one of the Fulbright Scholar academic sponsors (for author GJK) in Phnom Penh, approved this research (IR010; May 24, 2021). No patient records were accessed. No personal information was recorded. Data were collected from January 2021 through June 2024.

Respondents in rural Cambodian provinces identified needs comprising rapid on-time COVID-19 screening with high-quality COVID-19 antigen tests (top priority) that minimize false negatives, increased testing of people crossing borders and moving in and out of contagion zones, and increased accessibility of RT-PCR testing. They requested mobile testing for people living in immigration sites; at drive-throughs; in communities (families, villages, markets, and at pagodas); at emergency rooms; at nursing homes; and in airports. Several people noted the need for careful maintenance of biosafety, quality control, and supply chains. There was little discussion of antibody testing or the need for it. The responses allowed us to summarize technical diagnostic needs as follows.

Technical needs identified by the respondents included portable cartridge- or cassette-based molecular diagnostic instruments characterized with few processing steps to improve the ease of testing and assure biosafety in community settings, especially remote rural areas; adequate personal protective equipment and district hospital rapid response to accelerate treatment and preoperative diagnosis; operator training, quality monitoring, and performance evaluations; improved public education to encourage the use of POCT; specimen handling protocols in which onsite testing was not possible and specimens must be sent out; widespread self-testing to help mitigate outbreaks and stimulate economic recovery; and SARS-CoV-2 variant sequencing.

On the western border with Thailand, the three-province cluster of Banteay Meanchey, Battambang, and Siem Reap, plus the relatively isolated Koh Kong Province (identified in Figure 1), as well as Pre Veng and Takeo provinces on the eastern border with Vietnam, were named by the respondents as top priorities for public health intervention. Delta and then Omicron variants were detected in Thai transient workers in Battambang and other border regions. Oddar Meancheay (remote northwest) lacked testing resources despite increasing contagion. Other high-risk regions and their risk factors identified encompassed Kampong Cham (factories and urban spread); Kampong Speu (prison); Kandal (casinos, factories, prison, and urban spread); Mondulkiri; Phnom Penh (high caseload and lockdown districts); Sihanoukville (casinos); Stung Treng (remote northeast); Svey Rieng (factories and high contact-tracing alerts); and Tboung Khmom provinces.

Respondents emphasized the need for public release of testing data, clearer citizen instructions for mandatory testing upon entry to a province, communication of testing site locations, formal evaluation of imported test kits, and free RAgTs with easily interpreted results. One respondent noted that 3 to 5 days was too long to wait for RT-PCR results. Another said not to test vaccinated people unless symptomatic and recommended that RT-PCR testing be placed at or near borders to lower testing costs. All respondents wanted root-cause solutions that would decrease the spread of COVID-19 in areas of high population density.

Respondents stated that the goals of national policy and guidelines for COVID-19 should include public health strategies and recommendations for test-positive people and their families; systematic procedures for contact tracing, isolation, and self-quarantine; testing throughout Cambodia to document positivity rates; education regarding the differences in RAgTs, molecular diagnostics, and antibody tests; collaboration of bioengineers and informatics specialists; diagnostic criteria for advanced medical care and hospitalization; accessibility of diagnostics to vulnerable populations; and ultimately, accelerated evidence-based treatment of infected patients.

Figure 2 displays the Pareto plot for distance and time to testing in 17 border provinces (Table 1). A Pareto plot displays data in the order of descending magnitude along the horizontal axis. The legend identifies the provinces (and their locations in Figure 1). The median distance from province districts to testing was 36.5 (range 0.5-247) km, and the median transit time was 73 (1-494) minutes. Within provinces, the median of the paired differences in maximum versus minimum distances to testing was 70 km (P<.001); for transit time, this value is 2.3 hours, demonstrating uneven access to COVID-19 diagnostics.

Koh Kong Province (Figure 1), which has a long undeveloped coastline and a mountainous, forested, and largely inaccessible interior, had one of the longest distances to testing of 125 km and prolonged one-way transit time of 4.17 hours. A district in Stung Treng Province had the longest distance (247 km) and the most prolonged time to reach a testing site (8 hours, 14 minutes).

Figure 3 presents the Banteay Meanchey Province on the eastern border of Thailand. This province is divided into 9 districts with populations shown on the map. Table 2 details district names; main villages; and demographic, geographic, and risk features at the time of COVID-19 surge. During the 2021 peak, RAgTs were not available to the public. Among 72 public health facilities, 1 hospital offered RT-PCR testing (Cambodia-Japan Friendship Mongkul Borey Provincial Hospital). Two others (Serey Sorphorn and Poipet Referral Hospitals) offered RAgTs.

Figure 4 shows the Pareto plot for the Banteay Meanchey districts. The median distance to testing was 45 (range 5-75) km. In descending order, transit times ranged from 150 to 10 minutes. Phnum Srok District, an isolated mountainous area in which people face difficulties and delayed emergency care access in the northwest corner of the province, had the maximum distance (75 km) and transit time (2.5 hours; Table 2). Median transit time was 90 (range 10-150) minutes. The mean was 42.7 (SD 23.9) km.

Border provinces were locked down from July through August 2021 (Figure 1). A highway from Thailand passing through Banteay Meanchey Province into Siem Reap City, one of the largest cities in Cambodia, witnessed heavy traffic on normal days, which was not the scenario during lockdowns. The lockdowns isolated Siem Reap Province where the median time to a testing site was 51 (mean 85.8, SD 74.4, range 10-230) minutes. In general, in rural provinces, isolated lockdown regions had limited hospital beds, critical care facilities, respiratory ventilators, laboratory resources, diagnostic instruments, and personal protective equipment, which impaired health care access to treatment for rural populations.

Mobile and automated COVID-19 diagnostics in the rural Solano and Yolo counties of California presaged a framework of POC strategies and prehospital testing recommended for limited-resource settings in Cambodia. Coauthor AZ was the first to develop, implement, and manage COVID-19 testing in vans and transportable pop-ups starting in Vacaville (population 103,078), Solano County, in parallel with ongoing field research in Cambodia. Table S4 in Multimedia Appendix 1 summarizes test methods and operational features. Time to test result was 5 to 30 minutes. Patients received results immediately and through email. Vacaville was the site of the first community transmission of SARS-CoV-2 in the United States. Figure 4 in the study by Kost [32] presents the pictorial sequence of the first outbreak and the response from the Centers for Disease Control and Prevention. Mobile-testing vans in California were launched by CovidClinic, a nonprofit 501(c)(3) company that operated community testing at 78 locations in California and 140 clinics nationwide; in the discussion, we provide illustrations showing (1) the POCT design developed in the CovidClinic mobile laboratory and (2) the free-roaming-type clinic [42,43] operated by Optumserve. This roaming clinic provided COVID-19 onsite testing and prescription medications (Paxlovid and molnupiravir) via telehealth consultation for symptomatic patients who tested positive upon presentation at mobile clinic sites operated through February 2023 in Yolo County, which is adjacent to Solano County.

In Figures S1 and S2 in Multimedia Appendix 1, we compare RAgT test performances. In 2021, RAgT developed by Roche Diagnostics met the WHO performance criterion for sensitivity >80% (median) in clinical evaluations (Figure S1 in Multimedia Appendix 1), while PanBio RAgT did not (Figure S2 in Multimedia Appendix 1).

This research identified a lack of sufficient COVID-19 diagnostics, both RAgTs and molecular diagnostics, in isolated and rural areas of Cambodia during peak outbreaks, excessive transit time to testing sites in border provinces and high-risk regions where international crossings are located, statistically significant differences in access to testing within border provinces, and geospatial gaps in access to testing in noncentral provincial areas. Implementation of novel diagnostic portals, such as mobile testing and 24-7 automated test dispensers in Yolo and Solano counties of California, provided models for public health delivery that would be feasible to implement in other rural settings including the limited-resource border regions studied in Cambodia. High performance POCT strategies allow timely, accessible, and mobile diagnosis of asymptomatic and symptomatic patients at critical points of need.

COVID-19 demonstrated that weaknesses in health care small-world networks can generate adverse implications for vulnerable individuals, cultural welfare, workplace security, economic progress, and social instability [44-46]. Several thousand health care workers in the United States have died of COVID-19 [47]. The number of deaths in Cambodia is not known. Globally, the WHO estimated 115,500 COVID-19–related deaths among health care workers [48]. To protect medical professionals, first-time responders, oneself, families, and others from acute and protracted disease, especially to protect individuals older than 65 years (among whom 90% of COVID-19–related deaths were reported in November 2022 [49]), accessible COVID-19 diagnostics and rapid-test results remain paramount.

The COVID-19 pandemic inspired several mobile-testing solutions. In China, Xing et al [50] developed a mobile laboratory for onsite molecular diagnostics. Others deployed inflatable molecular diagnostic laboratories in Beijing [51]. In Singapore, Linster et al [52] published a mobile design and highlighted its biocontainment features. In Victoria, Australia, Ballard et al [53] described a testing van with ISO 15189 certification and illustrated testing process steps (see Figure 1 in the study by Ballard et al [53]). In addition, in Australia, Paton et al [54], with the assistance of Lab Without Walls, deployed mobile molecular diagnostics for the detection of SARS-CoV-2.

In Angola, Africa, Owens et al [55] devised a mobile laboratory curriculum. In the United States, Kost et al [56] created a comprehensive POCT educational curriculum for public health learners. In South Korea, Roh et al [57] compiled guidelines for mobile testing comprising installation, equipment, evaluation, sample processing or storage, testing or reporting, internal or external quality control, management, personnel, infection control, biosafety, and information systems. Affara et al [58] described mobile laboratory networks in Africa. In East African countries, Gehre et al [59] showed that mobile testing providing certifications of uninfected drivers helped facilitate business enterprise when borders closed. Tong et al [60] used spatial modeling to help control COVID-19 in Hong Kong. However, extensive literature searches did not recover any reports of COVID-19 mobile-testing laboratories or geospatial analysis of testing sites in Cambodia.

Mobile-testing vans provide diagnostic portals [10], multiplex testing, professional personnel, quality assurance, and telehealth. Accessible diagnostics (Figure 5) in Solano and Yolo counties in California fulfilled public health needs for distributed COVID-19 testing and surveillance in these rural regions. From July 2022 to February 2023, a total of 51,000 RAgT kits were given out [61]. The mobile clinic with telehealth consultation started operating in Yolo County during the fall surge of 2022 as one element of the California “Test-to-Treat” program. Free masks (ie, N95) can be added to these accessible resources, as well as other public health supplies, such as mosquito repellents, sunscreens, condoms, pregnancy tests, sexually transmitted infection tests, and overdose reversal drugs such as naloxone [61]. Personal protective equipment must be worn in mobile-testing zones, which are considered active laboratories, vans must maintain appropriate ambient environmental temperature and biosafety measures, and in the United States the Occupational Safety and Health Administration requires fire escape plans.

Of Cambodia’s 17.2 million people, approximately 80% live in rural areas in which mobile services can improve health care. Transit time, return trips, and lost wages diminish motivation to make arduous trips to referral hospitals for testing [62]. Rural communities have a higher proportion of older adults than urban communities; the frequency of noncommunicable diseases is higher among the rural poor [63-66]. Migrant workers crossing borders spread COVID-19. Sparsely placed rural health services and infrastructure made it difficult to effectively respond to COVID-19, its variants, and confounding diseases [67], which, if neglected, can result in excess mortality.

Table 3 summarizes challenges encountered when implementing mobile van testing and lists recommendations for implementation of POCT in rural and limited-resource settings. In early May 2021, during the height of the COVID-19 surge in Cambodia, we recommended [62] free distribution of RAgTs for self-testing throughout villages and mobile RAgTs and RT-PCR testing at key locations (indicated by the red dots in Figure 3) including border check points, to identify infections where people crossed borders to reach markets, work, and casinos. Generally, one-way distances and travel times from Banteay Meanchey villages to RT-PCR testing were too prolonged (median 45 km; Figure 4) to motivate travel to molecular diagnostic testing sites.

Awareness of infection through COVID-19 testing can help meet several identified needs. Transmission could have been slowed by checkpoint and self-testing to detect COVID-19 variants such as Delta in 2021, and later Omicron, which spread quickly across Thai-Cambodia borders. Unfortunately, the costs (US $2-$7) of most RAgTs were prohibitive for rural Cambodians at the time of 2021 peak surges. Test kits were not generally available nor distributed freely in rural areas. Eventually, the Cambodian government approved and distributed limited supplies of RAgTs, which we encouraged through communications, lectures, and webinars with the National Public Health Laboratory staff, university, and professional participants.

Rapid self-testing and timely awareness can help people adapt to endemic diseases that demand long-term medical and social change [68]. Empowered with immediate test results, people can receive guided treatment using telehealth to optimize diagnostic-therapeutic cycles [69]. Geospatial strategies in limited-resource settings should address “spatial justice,” that is, “the fair and equitable distribution in space of socially valued resources and the opportunities to use them” [70]. Home, community, and emergency testing [10,40], including layperson self-testing and testing among families at an affordable cost or subsidized cost by government agencies can help establish spatial justice for rapid diagnosis. Benefits of bedside, mobile, door-to-door, and prehospital testing comprise high-volume testing; improved access; privacy; multiple test formats; service for hard-to-reach, marginalized, vulnerable, and underserved people; diminished disparities and inequities in mortality rates; tracking of test positivity rates; outbreak mitigation; fulfilling needs in isolated island and regional rural settings [71-81]; and improving patient outcomes [82-84].

For high-risk border provinces and rural outbreaks, disparities in access justify distributing and mobilizing testing. Even now, quick access to COVID-19 and other infectious disease tests could help avoid older adult deaths and the spread of pediatric infections. A sustainable business model might be enabled by including tests for infectious disease surges, such as seasonal influenza A and B (2022, highest cases in over a decade), respiratory syncytial virus with portable multiplex molecular assays, and increasingly prevalent community infections such as strep throat [85]. In rural communities threatened by de novo pathogen transmission, multipurpose mobile testing, automated test kit access, low-cost diagnostics, data-driven surveillance systems [86], and telehealth treatment today seem a rational way to prepare for tomorrow. Crisis standards of care [87,88] would help prepare, codify, sustain, and enhance the use of novel POC strategies in anticipation of the next pandemic.

This research successfully documented geospatial constraints in access to COVID-19 testing in rural Cambodian border provinces. Early POC diagnosis helps find new pathogens, limits their spread, and informs public health of increasing threats [89]. However, limited scope and government restrictions prevented the discovery of how outcomes were affected by changes in spatial testing patterns during and following COVID-19 surges in 2021. The Cambodian government approved the use of RAgTs in private hospitals and clinics in late April 2021 [90] and, on July 7, 2021, approved RAgT self-testing in homes with Roche, PanBio, Standard Q, and Indicaid kits [91]. Although the Ministry of Health published a RAgT operating procedure for border gates, factories, and other business locations in July 2021 [92], the testing rate in Cambodia was too slow to keep pace with the spread of new viral variants (eg, Delta in 2021 and later, Omicron).

Then, in February 2022, the government declared it a crime if citizens withheld positive COVID-19 test results [93]. On October 1, 2021, the government stopped administering RAgTs and ceased reporting RAgT-positives; case numbers reported fell 76% [94]. Consequently, by March 2023, Cambodia ranked 125th worldwide for COVID-19 cases (138,719) with 3056 reported deaths (104th), 94th for total tests (3,091,420), but 162nd for tests per 1 million population (180,062/million) [11]. Notwithstanding government policies, RAgTs cannot reliably rule out COVID-19. The US Food and Drug Administration has approved home tests for simultaneous detection of COVID-19 and influenza [95]. However, home molecular diagnostics (eg, loop‐mediated isothermal amplification), which demonstrates higher levels of performance for ruling out and ruling in infectious diseases, should be taught in public health curricula [96,97] and set expectations for future home testing [98,99]. To mitigate these limitations, we also recommend that national guidelines for high performance and cost-effective diagnostics [100] be put in place proactively in limited-resource settings to anticipate emerging variants, new outbreaks, and a possible future pandemic, as well as to improve the ongoing accessibility of COVID-19 and other vital diagnostics for highly vulnerable populations, such as people living with HIV [101].

The COVID-19 crisis propelled POCT directly into communities to meet the needs of people for quick access to testing in homes and at nearby diagnostic portals, although as noted earlier, local prevalence and the presence of symptoms influence RAgT reliability. Experience with rural American mobile testing and automated test kit dispensing suggests that these approaches could be implemented in limited-resource rural provinces in Cambodia, in remote settings of other countries, and in island nations to build public health resilience. Hence, public health practitioners should be prepared to implement POCT strategically in communities and homes, enable prehospital mobile testing, provide automated test dispensers, establish telehealth consultation, and expedite treatment when indicated by positive diagnostic results obtained rapidly at points of need.

Automated vending machines that dispense COVID-19 RAgTs and home molecular diagnostics kits (eg, loop‐mediated isothermal amplification tests) 24-7 could be positioned in public health vehicles and POCT vans as shown in Figure 5, facilitating safe testing and reliable detection of SARS-CoV-2 or new emerging viruses and variants in hotspots in which symptoms are experienced first. Targeted vaccination, follow-up community testing, travel certificates, telecommunicated epidemiological warnings, and the addition of public health data to surveillance repositories represent additional potential benefits. Mobility could be extended to wellness testing to cost-effectively enhance public health outreach in underserved areas.

Quick detection, control, and mitigation of outbreaks require adaptable, flexible, and dynamic responses that fit geographic, social, environmental, seasonal, and cultural expectations. When not responding to infectious diseases or other crises, mobile and prehospital POCT could enable rapid response care close to homes in rural and remote areas while playing an important role in educating health care providers about POCT field practice and its advantages. Public health facilities, budgets, and capabilities differ from country to country. Therefore, national POCT guidelines should be written in anticipation of infectious disease crises to accelerate regional responses and their geospatial optimization.

RAgTs are more effective when symptoms and viral load peak. Testing within communities at optimal sites determined by quantitative geospatial analysis would help close gaps in public health accessibility, reveal health care discrepancies, and speed diagnostic test results in limited-resource areas. Geospatial planning can enhance resilience, assure spatial justice, and deliver diagnostic testing to highly vulnerable populations such as people living with HIV. Public health assets should be characterized by higher-performance, lower-cost, readily accessible, and user-friendly POC technologies such as molecular diagnostics for self-testing and distributed border detection for public health surveillance.