Study design and population

This study was a cross-sectional survey using pilot-tested questionnaires conducted over the phone by trained volunteers. We included those PHCs registered in the MoPH network in both years (2019 and 2020) that had also agreed to share their data with a third party. We included all patients or patient caregivers who were at least 18 years old and who had received a health care service in 2019 or 2020 from one of the sampled PHCs. We excluded patients with no available contact information.

The full list of staff working in the PHCN was provided by the MoPH. From that list, we included employees who were clinical staff providing the basic PHC services as per the national PHC standards (general physician/family physician, pediatrician, gynecologist, cardiologist/endocrinologist, dentist, and nurses) or administrative/managerial staff. We included only employees who were employed in the sampled PHC in both years (2019 and 2020) in order to describe the potential changes in the provision of essential health care services in the PHCN, taking the year 2019 as a reference. We excluded staff who worked in the PHC in only one of the years (2019 or 2020).

Sampling

We followed a two-stage cluster sampling method. We first selected PHCs in the PHCN (i.e., clusters); and then selected staff and patients from each of the included PHCs. To select the PHCs, we used systematic probability proportional to size sampling relying on the number of consultations in each PHC in 2019 as the size measure. The sample size was 96 PHCs (out of a total of 175 PHCs) for a 95% confidence interval (CI), assuming a conservative 50% expected proportion of PHCs affected by services disruption and 10% precision.

For the patients, we estimated the sample size to be 768 patients based on a 95% CI, an estimated conservative proportion of 50% experiencing service change, with a precision of 5% and accounting for the cluster effect. Using simple randomization and accounting for a 30% expected response rate (based on the pilot and unpublished surveys by the PHC team in the MoPH), we selected 2592 patients from the patient lists for each PHC (27 patients per PHC).

For the staff, we estimated the sample size to be 192 staff taking into account an expected conservative proportion of 50% of staff who would perceive that the utilization of services was affected, with 95% CI and 10% precision. Using simple randomization and accounting for an expected response rate of 35% (based on the pilot and unpublished surveys by the PHC team in the MoPH), we selected 576 staff from the list of staff (6 staff per PHC).

We calculated the needed sample sizes using Epi Info version 7.1.2.0 (U.S. Center for Diseases Control and Prevention). The sample selection of the PHCs, patients, and staff was performed by the MoPH information technology (IT) administrator with the guidance of the research team.

Data collection

The data were collected using pilot-tested questionnaires. The questionnaires were developed for the purpose of this study first in English, then translated into Arabic, and then back translated into English. They were administered to the interviewees (patients and staff) in Arabic.

The patient questionnaire included questions on socio-demographic characteristics; utilization of services; adoption of COVID-19 infection prevention and control measures in the PHC; and the perceived role of PHCs in response to COVID-19. The staff questionnaire included questions on socio-demographic characteristics; difficulties encountered with service continuity; measures adopted in the PHC to protect the staff, patients, and the community against COVID-19; and the perceived role of PHCs in the pandemic.

The data collection was conducted from June to July 2021 by Lebanese Red Cross volunteers. The teams underwent one-day training on research ethics (delivered by the first, fourth and seventh authors of this study) focusing on obtaining informed oral consent and on administering the questionnaires. The questionnaires were administered over the phone and the answers were directly recorded into Survey123 (ArcGIS, Esri, United States of America). The teams were supervised and supported by one or more staff from the International Committee of the Red Cross (ICRC) to address any inquiry or issue.

Ethical considerations

This research project was reviewed and approved by the ethical review board of London School of Hygiene and Tropical Medicine (Ref: 22760), by the International Committee of the Red Cross (ICRC) Health Unit in Geneva, and by an independent group of experts within the Lebanese Ministry of Public Health.

Prior to participating in the study, oral consent forms were obtained from all participants. The consent process involved providing detailed information about the study’s objectives, procedures, potential risks and benefits, and the rights of the participants. Participants were given the opportunity to ask questions and clarify any concerns before providing their oral consent. It was explicitly communicated to the participants that their data would be treated with strict confidentiality, and they were informed of their right to withdraw from the study at any time without any repercussions.

This study adhered to the principles of research ethics, including the respect for autonomy, confidentiality, and the protection of participants’ rights. The study protocol complied with the guidelines outlined in the Declaration of Helsinki for research involving human subjects.

Data analysis

We reported proportions and percentages for categorical variables (e.g. gender, nationality) and means ± standard deviations (SDs) for continuous variables (e.g. age, hours in PHC per week). We used the Chi-square test, Fisher’s exact test and the student’s t-test when applicable. We conducted multivariable analysis using the Generalized Estimating Equations (GEE) appropriate for correlated data and stated objectives. We identified the factors associated with the dependent variables: change in utilization of services from the patients’ perspective and change in provision of services from the staff’s perspective using ‘no change’ as the reference category. We included the characteristics of the patients and those of the staff (e.g., age, gender, and nationality) as independent factors. We reported crude and adjusted odds ratios (ORs) with 95% confidence intervals (CIs) and set a p-value < 0.05 for statistical significance. We used IBM SPSS Statistics for Windows, Version 26.0 (IBM Corp., Armonk, NY, USA) for statistical analysis.

Description of the included population

We contacted 1600 patients and 375 staff distributed across 96 PHCs in Lebanon (Supplementary material 1 in S1 File). Out of those contacted, 508 patients and 89 staff (32% and 24%, respectively) were not reachable, and once contacted, 194 patients and 19 staff (12% and 5%, respectively) were not eligible. From those who were reachable and eligible, 782 patients (87%) and 214 staff (80%) consented to be surveyed. Consequently, the response rate was 49% (n = 782, N = 1600) for patients and 57% (n = 214, N = 375) for staff. We excluded 19 patients and 16 staff due to incomplete responses, resulting in a total of 763 patients and 198 staff completing the surveys and their data analyzed. There was no statistical difference between the total sample and the participants reached in terms of age, gender or nationality. Additional details can be found in Fig 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Flowchart of the patient and staff populations included in the study.

https://doi.org/10.1371/journal.pone.0288387.g001

Patients

The mean age of these 763 patients was 40 years (SD = 12.4) and more than half were female (n = 425, 57%). Most of the patients were Lebanese (63%), with the non-Lebanese patients (37%) mainly including Syrian patients (98%, n = 277). More than half of the included patients were caregivers (56%). Of the total population surveyed, the caregivers were in general caring for children (80%) rather than elderly. The patients attended mainly PHCs in Mount Lebanon (42%) and attended for both years 2019 and 2020 (66%). More than half of the patients (58%) and 28% of primary earners were unemployed. Additional details on patient characteristics are presented in Table 1. Statistically significant differences were seen between the socio-demographic profiles of Lebanese and non-Lebanese patients (See Supplementary material 2 in S1 File).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Characteristics of the patients included (N = 763).

https://doi.org/10.1371/journal.pone.0288387.t001

Staff

Among the total of 198 staff, the mean age was 40 years (SD = 12.6) and more than half were female (54%). Almost all the staff were Lebanese (99%) and worked in PHCs across Lebanon, mostly in Mount Lebanon (41%). The mean number of working hours in the PHC was 22.8 hours per week (SD = 18.5). Staff represented different occupations within PHC, including but not limited to nursing (26%) and general practice/family medicine (22%). Additional details on staff characteristics are presented in Table 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Characteristics of the included staff (N = 198).

https://doi.org/10.1371/journal.pone.0288387.t002

Reported changes in utilization and provision of essential health care services

Patients.

The majority of the patients interviewed attended the PHCN in both years 2019 and 2020 (n = 501, 66%) and the remaining attended the PHCN in one year only, either 2019 (n = 117, 15%) or 2020 (n = 145, 19%). There was no statistically significant difference between Lebanese and non-Lebanese patients in the proportion of patients attending the PHCN in 2019, 2020 or both years (crude OR = 1.606, 95% CI [0.960, 2.688]). Of the patients who attended the PHCN only in the year 2019, almost one third declared during the interview they would have needed to (but did not) seek multiple services from the PHCN in 2020 (n = 29, 28%), with a higher proportion of non-Lebanese (n = 14, 44%) compared to Lebanese patients (n = 15, 21%) (crude OR = 2.852, 95% CI [1.157, 7.029]). The most needed service would have been immunization (n = 17, 59%) (Supplementary material 3 in S1 File).

Interrupted services were reported by 15% of all patients (94 out of 635) who used services in 2020 (either in 2020 only or in both 2019 and 2020). Out of these 94 patients, 73 (78%) specified which services were interrupted. The main services that were interrupted were access to chronic (n = 49, 67%) and acute (n = 29, 40%) medication. Specifically, the disrupted access to chronic medications was considered as the ‘most important interrupted service’ by 65% (n = 47/73) of the respondents who specified the interrupted service. The PHCs most affected by service interruptions were in Akkar (n = 31, 33%), North Lebanon (n = 28, 29%) and Mount Lebanon (n = 23, 25%) (p-value = 0.001).

The multivariable analysis included a total of 548 patients (Table 3). Older patients had higher odds of having interruption of services (adjusted OR = 1.044, 95% CI [1.019, 1.069]). Patients of PHCs from Akkar governorate had higher odds of interruptions compared to those from Beirut governorate (adjusted OR = 4.831, 95% CI [1.353, 17.246]), while patients with high school degrees had higher odds of interruptions compared to those with no education (adjusted OR = 2.829, 95% CI [1.188, 6.735],). No other statistically significant factors were identified.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Factors associated with the interruption of services from the patients’ perspective.

https://doi.org/10.1371/journal.pone.0288387.t003

Staff.

One third of the staff reported interruptions in the provision of services (n = 64, 33%). The multivariable analysis included a total of 174 staff (Table 4). Having a gynecologist present was protective against interruption of services (adjusted OR = 0.074, 95% CI [0.01, 0.560]); this was the only statistically significant factor associated with service interruption.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Factors associated with the interruption of services from the staff’s perspective.

https://doi.org/10.1371/journal.pone.0288387.t004

Reasons reported for changes in utilization and provision of essential health care services

Perception on implementation of measures for infection prevention and control

Perceptions related to the role of the PHCN during the COVID-19 pandemic

Summary of findings

In the first months after the onset of the COVID-19 pandemic in Lebanon, there was a continuity in utilization of PHC services from the patients’ side, as well as a continuity in the offer of essential services from the PHCN. The services that were most negatively affected during that period were those related to immunization and NCDs. However, it seems that the disruption was provoked mainly by economic barriers rather than by the implementation of lockdown measures. This finding might explain why the Lebanese population, rendered increasingly vulnerable from multidimensional poverty, turned to the public network to access essential health care services.

The perceived quality of services offered by the health workforce and received by the patients was seen as acceptable with regards to the implementation of IPC measures. However, patients attending the PHCN reported receiving less health information and communication than what was reported delivered by the staff.

Both patients and the health workforce considered that the PHCN could–and should–have played a larger role in the country’s response to the pandemic.

As a new epidemic is today hitting the country, with cholera spreading across the most vulnerable communities, starting once again in the poorest areas in Akkar and the North, the findings from this study might be of critical relevance in informing the ongoing response to the multifaceted humanitarian crisis now unfolding in Lebanon [18].

Interpretation of findings

The population participating in our study seem to confirm that the service users of the Lebanese PHCN are the most vulnerable groups: in fact, more than one third (37%) of study participants were refugees, and over half (58.1%) were unemployed. Such findings underline the importance of the PHCN as a safety net in a situation of fragility and support its efforts to ensure UHC [16]. Our findings on the geographic dimension of vulnerability in Lebanon are corroborated by other studies conducted in the country under the lead of United Nations Office for the Coordination of Humanitarian affairs [21], which identify Akkar and the North Governorates as the most vulnerable regions.

Despite 2020 being the year marking the beginning of the COVID-19 pandemic, these months saw a marginal increase in the number of patients attending the PHCN in Lebanon compared to 2019 (145 new patients). The overall services interruption experienced by patients was lower than what we had conservatively assumed based on data available from the country (i.e., 15% versus 50%) [19, 22]. The workforce of the PHCN had a perception of higher interruption of services (33%) compared to those who used the services, probably due to the increased difficulties the providers faced in trying to ensure continuity of service provision [23].

In terms of specific type of services, NCD care emerged as the most severely interrupted, as has been reported in several regions of the world [1, 24]. Immunization was also mentioned as an important forgone service. This is consistent with other studies conducted in Lebanon [19], as well as with findings from the global literature [25].

What emerged as elements supporting continuity of PHC service delivery in the PHCN were not only the greater affordability of services but also a reported increased knowledge of and trust in services available. This element of increased awareness and knowledge of services available testifies to the learning journey of the MoPH PHC Department since the beginning of the Syrian crisis and its spillover effects on Lebanon [26]. The PHCN has sought the opportunity- throughout the refugee crisis and the subsequent shocks that have put it under strain- to strengthen its capacity to respond to the needs of vulnerable populations in the country and to address what had been highlighted as key challenges for them in promoting utilization of essential health care services [17, 27].

Our study highlights how the COVID-19 pandemic, associated in time with the additional shocks affecting Lebanon from 2019 onwards, has hit particularly hard on the most disadvantaged populations in Lebanon. In fact, the vast majority of reported interruptions in the utilization of PHC services was reported by the non-Lebanese population in the country. This finding adds to the body of evidence that compounding shocks often have an inequitable distribution, as they disproportionately affect refugee communities [28, 29].

The several crises that have beset Lebanon in the last five years have had a cumulative impact on the continuity of core health services. It is difficult to unpack the effect of any single shock. Health systems are increasingly being described as complex adaptive systems [30], and in a setting affected by protracted crises, such as Lebanon, it might prove particularly relevant to adopt a resilience lens in analyzing the responses to compounding stressors [31]. Particularly in fragile and conflict-affected states, key dimensions of resilience within health systems are found in the assessment of interdependencies among relevant economic, socio-political, and epidemiological factors as well as in analysis of the overall uncertainties generated by the deepening volatility of society with associated greater vulnerability of the affected populations [32]. The intersections between epidemics and poverty have been well described in other geographical settings [33, 34], but have been poorly documented in the Middle East. More in-depth studies are warranted looking at the systems dynamics unfolding in this specific region. In particular, it would be useful to adopt methodological approaches based on systems thinking to unpack the complexity of the ecosystem within which health systems operate [35].

Interestingly, pharmacies emerge as an ancillary coping mechanism for patients who seek to access drugs and to obtain health information. According to the latest survey conducted, there are 2968 community pharmacies in Lebanon, which represents a significantly broader and more easily accessible network for patients than the PHCN, particularly in remote areas [36]. It has been well documented how, throughout the pandemic, pharmacists have often turned to be frontliners in the response, emerging as the preferential entry point into care for patients who could not afford other services [37–39]. However, there has been little engagement with them from MoPH as well as from the humanitarian sector. Our findings might be a case in point for considering integrating community pharmacies in the PHCN, as has already been advocated [40].

Patients reported good adherence of PHCs to IPC measures, particularly with regards to wearing masks (99%), physical distancing (95%), and temperature control (93%), while they reported less satisfaction with education and information received (70%). In particular, non-Lebanese patients reported lower satisfaction with education and information sessions compared to Lebanese patients. These findings appear to be confirmed also beyond COVID-19 services: another study conducted in Lebanon among NCD patients highlighted the disparity in access to services and education. When comparing the provision of health services by nationality, Syrian refugees were less likely to receive the necessary services and lacked awareness on the management of their conditions [29].

The perception of the staff with regards to the implementation of IPC measures was different than that of patients. In fact, the staff reported lower availability of PPE and slightly higher levels of awareness, education and information provided to the communities. This divergence in reports is a point that warrants further research, considering the conflicting reports in the literature with regards to adherence to IPC measures among physicians in Lebanon [41].

Despite the lack of essential diagnostics and the poor options for risk communication available, patients experienced the PHCN as more accessible and affordable than other types of services. Both patients and staff believed that the PHCN should have played a larger role in responding to the pandemic, strengthening the need for a greater role for PHCs in epidemic preparedness and response, as already advocated for in Lebanon [20].

The key role of the PHCN in terms of proximity and understanding the needs of the patients makes it the ideal candidate not only for early warning in the countryside and provision of health security at country level, but also more broadly for UHC. Despite the development of early warning systems in the Middle East, it has in fact been highlighted how their performance remains substantially substandard in terms of population coverage [42]. In the context of the fragmented Lebanese PHC system, effectiveness of epidemic surveillance is jeopardised by the weak integration of the private network into the MoPH surveillance system. It is therefore imperative to take stock of these findings particularly in this historical moment, where Lebanon is facing a cholera epidemic which is adding more strain to an already overburdened health system [18]. Investing in the PHCN now, at this pivotal time, might help decrease the burden on hospitals as well as provide multiple entry points into prevention and care for the most vulnerable, supporting the efforts of the PHCN to progress towards UHC.

Strengths and limitations

Our study is the first cross-sectional study of its kind, exploring changes in utilization and provision of services in the PHCN in Lebanon through a representative sample of both service users and service providers. The fact that both patients and staff were included in the study population provides a unique opportunity to document perceived reasons for changes from multiple viewpoints; and to triangulate our findings with what has been documented in previous studies analysing only specific subgroups of service users and providers, or specific essential health care services.

There are, however, some limitations. In particular, the very long period covered by our inquiry (two years), might have inevitably led to an important recall bias. Further, multiple factors have been investigated simultaneously, which might add an unknowable element of confounding that could not be controlled for in the analysis. A second element that needs to be taken into account is the fact that the public PHCN represents approximately only 20% of the total number of facilities providing primary care at country level, with its 254 PHCs out of a complex and diverse landscape of over 1000 outlets across the country, established by private profit and non-for-profit actors who did not systematically provide routine health services data to the MoPH [17]. Therefore, the findings here might not be generalizable to the whole population. We consider them still relevant nevertheless, as this information is systematically drawn from the most vulnerable populations in country who are unlikely to have access to non subsidised care and are therefore more likely to be the main users of the public network. This study provides insight into the barriers to accessing primary care for those who are the most at risk of being left behind, with the added value of identifying barriers that are specific to them.

The consistency between the findings of our study and those reported earlier in the literature corroborates the validity of our results, although the study was not powered for the substantially smaller change in utilisation of services documented compared to what was conservatively expected (i.e. 15% versus 50%). There was no evidence, before we analyzed our results, pointing towards such a small change in utilisation, as the literature available indicated a much greater negative impact of COVID-19 on utilisation of essential health care services in Lebanon [19, 25]. Despite the discrepancy between what was expected and what evolved in terms of response rate, the findings themselves that emerged are consistent these other studies and corroborate the hypothesis that specific types of services were particularly negatively affected—namely immunization and NCD care [19, 29].

It is also the case that the ultimate sample used in this study was less than half of the original sample sought in our initial sampling frame. Reasons for the progressive exodus of patients and PHCs from our study are briefly mentioned (not reachable, not eligible, reluctance to participate). It is possible that there are underlying factors which may have skewed our final sample in unknown ways. To address this concern, a subsequent analysis of our initial outreach sample by age, sex, and nationality showed marked similarity in demographic parameters with the sample reached for our study.

Implications for public health practice and research

Our study sheds light on the critical role of the PHCN in Lebanon as the last resort provider in times of hardship for the most vulnerable populations in the country. It is therefore imperative that donors and the aid community join efforts in ensuring additional support to strengthen this key contribution, particularly in this critical moment of intensifying economic crisis, collapse of essential services such as electricity and sanitation, removal of subsidies for essential drugs, and an ongoing cholera outbreak.

The PHCN has proven resilient to the cascading shocks affecting the country and has learned lessons from the past. Engagement with communities and stronger risk communication emerge as an area of further improvement that will need additional support from the international community.

The emerging role of pharmacies as frontline providers warrants further investigation, as it might be an element to consider in the development of the PHCN as it seeks to address the growing humanitarian needs of vulnerable Lebanese and refugee populations in the country.