Response to the coronavirus disease 2019 (COVID-19) pandemic at private retail pharmacies in Kenya: a mixed methods study

Background: Private retail pharmacies in developing countries present a unique channel for COVID-19 prevention. We assessed the response to the COVID-19 pandemic by pharmacies in Kenya, aiming to identify strategies for maximising their contribution to the national response. Methods: We conducted a prospective mixed-methods study, consisting of a questionnaire survey (n=195), a simulated client survey (n=103), and in-depth interviews (n=18). Data collection started approximately seven months after the pandemic reached Kenya. Quantitative data were summarized using measures of central tendency and multivariable modelling done using logistic regression. Qualitative analysis followed a thematic approach. Results: The initial weeks of the pandemic were characterized by fear and panic among service providers and a surge in client flow. Subsequently, 61% of pharmacies experienced a dip in demand to below pre-pandemic levels and 31% reported challenges with unavailability, high price, and poor-quality of products. Almost all pharmacies were actively providing preventive materials and therapies; educating clients on prevention measures; counselling anxious clients; and handling and referring suspect cases. Fifty-nine pharmacies (55% [95% CI 45-65%]) reported ever receiving a client asking for COVID-19 testing and a similar proportion supported pharmacy-based testing. For treatment, most pharmacies (71%) recommended alternative therapies and nutritional supplements such as vitamin C; only 27% recommended conventional therapies such as antibiotics. While 48% had at least one staff member trained on COVID-19, a general feeling of disconnection from the national program prevailed. Conclusions: Private pharmacies in Kenya were actively contributing to the COVID-19 response, but more deliberate engagement, support and linkages are required. Notably, there is an urgent need to develop guidelines for pharmacy-based COVID-19 testing, a service that is clearly needed and which could greatly increase test coverage. Roll-out of this and other pharmacy-based COVID-19 programs should be accompanied with implementation research in order to inform current and future pandemic responses.


Introduction
The coronavirus disease 2019 (COVID- 19) pandemic is an ongoing global health emergency [1,2]. As of September 2021, over 230 million cases had been confirmed globally resulting in over 4.7 million deaths [3]. For the better part of 2020, non-pharmaceutical interventions such as early case detection, contact reduction, and hygiene measures were the main prevention strategies [4,5]. In early 2021, following unprecedented public-private collaboration and speed in research and development, a number of highly effective vaccines became available [6,7].
Non-pharmaceutical supportive care remains the key approach to treatment and management [8][9][10]. Among hospitalized patients, pharmaceuticals that have been shown to be effective include corticosteroids such as dexamethasone, antivirals such as remdesivir, and immunomodulators such as tocilizumab [8,11,12]. In the outpatient setting, pharmaceuticals used to manage mild COVID-19 cases include those targeting fever and pain such as ibuprofen, and those aiming to relieve respiratory congestion [9] [12]. Other more specific agents that have been tried, such as hydroxychloroquine, ivermectin, and azithromycin, are discouraged [9,12]. The World Health Organization and the US National Institutes of Health recommend that such unproven therapies be used only within the context of clinical trials [13,14].
It is predicted that follow-up waves of COVID-19 infection may continue indefinitely [15]. Strategies are therefore required to enhance and sustain coverage, cost-effectiveness, and efficiency of prevention and treatment interventions. Private retail pharmacies (also referred to as community pharmacies) are a unique channel through which COVID-19 prevention and treatment interventions can be delivered, since they are often the first or only point of contact with the healthcare system in developing countries [16][17][18]. The main reasons for preferential care-seeking at pharmacies, compared to health facilities, include greater accessibility, lower cost, and greater perceived privacy [19,20]. The International Pharmaceutical Federation highlights key areas where the pharmacy sector can support the pandemic response, including: management of the medication supply chain, patient education, identification and referral of patients to other health care providers, and provision of prevention materials and therapies [21,22]. During the COVID-19 pandemic, pharmacies in many settings, including Kenya, were categorized as essential services hence they were exempted from closure [23].
The first COVID-19 case in Kenya was reported on 13 th March 2020, and as of 12 th October 2021, there were 251,248 confirmed cases and 5,190 deaths [24]. A number of control measures have been instituted by the government at different times depending on rates of infection and mortality [23,25,26]. Vaccine roll-out in Kenya started in March 2021 [27]. We assessed the preparedness and response to the COVID-19 pandemic at private retail pharmacies in Kenya, aiming to identify strategies for maximising their contribution to the national response.

Study design
We conducted a cross-sectional descriptive study, including a questionnaire survey, a simulated client survey, and in-depth interviews. A mixed-methods approach was used in data collection and analysis: insights from interviews were used to refine the survey tools and probe quantitative data; and survey data were used to refine the interview guide and qualitative data analysis. Given the physical distancing requirements at the time, data collection was done remotely.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021. 10.22.21265188 doi: medRxiv preprint We targeted pharmacies that had participated in a pharmacy HIV prevention (PHP) study about a year previously (manuscript in preparation). Pharmacies had been selected randomly from the list of licensed pharmacies. Respondents were pharmacists (degree holders) and pharmaceutical technologists (diploma holders). Interview participants were purposively selected from among questionnaire respondents, targeting those that showed enthusiasm for the study through prompt and detailed responses. We aimed to increase variability of interview data by stratifying the sample by county, practice setting (urban vs. rural) and gender (male vs female).

Data collection
Data collection started in November 2020 (approximately seven months after the pandemic reached Kenya) and was concluded in December 2020 (for the surveys) and April 2021 (for the interviews).
One structured online questionnaire, taking about 20-30 minutes, was completed by each participating pharmacy (Supplementary file A). Data on pharmacy characteristics were imported from the PHP study database.
Simulated clients (SCs) made unscheduled phone calls to participating pharmacies, mimicking a client seeking COVID-19 preventive therapies (Supplementary file B). The SCs (2 women and 2 men, aged 25-35 years) had previously been engaged in the PHP study and trained comprehensively over a 2-week period. For the current study, further training was provided over a 1-week period through the Teams® digital meeting platform. An online debrief questionnaire was used to record observations (Supplementary file C). All calls were made on weekdays (Monday -Friday) between 8am and 5pm, over a 3-week period. Median [range] duration of calls was 4 [1][2][3][4][5][6][7][8][9][10][11][12] minutes. The majority (68%) of calls were made to a publicly available phone number, while the rest used a number in the study database as there was no public number. A few weeks after the simulated call, participating pharmacies were contacted to check if they suspected any clients visiting the pharmacies during the survey period to be "fake" (detection survey).
In-depth interviews were conducted by the first and second authors using a semi-structured interview guide (Supplementary file D). Interviews were conducted via the Teams® digital meeting platform as a first option or a simple phone call if the participant so preferred or if internet connectivity was poor.

Data management and analysis
Quantitative data were captured directly into an open-source database (REDcap®, Vanderbilt University). Data cleaning and analysis was carried out using Stata® version 15 (College Station, Texas, USA). Summary statistics were compared across counties and multivariable modelling . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.22.21265188 doi: medRxiv preprint done using logistic regression. Systematicity in detection was assessed by correlating detection status with caller, pharmacy characteristics, and simulation process features.
Qualitative interviews were audio-recorded and transcribed verbatim in Microsoft Word®, translating into English where necessary. To ensure quality, all transcripts were cross-checked against the audio recordings. Transcripts were imported into NVivo 12 (SR International, Australia) and coded by the first and second authors. Analysis followed a thematic approach with initial categories based on the interview guide and emergent themes integrated in a second round of coding. A checklist on Consolidated criteria for Reporting Qualitative research (COREQ) was completed (Supplementary file E).

Ethical considerations
The study was approved by the Kenya Medical Research Institute (KEMRI) Scientific and Ethics Review Unit (KEMRI/ SERU/CGMR-C/208/4080). Participants signed written informed consent forms administered digitally.

Response rate and characteristics of participants
Of 195 target pharmacies, 108 (55%) completed a questionnaire. Participating pharmacies were generally similar to non-participating pharmacies, but had features indicative of higher operational capacity and interest in public health interventions, such as having a computerized stock management system, written job descriptions, and providing blood pressure measurement (data not shown). Ninety-four (87%) pharmacies were located in urban areas; 44 (41%) had a consultation room; 71 (67%) had a computerised stock management system; and 13 (12%) were part of a chain network (Supplementary table I). The majority of questionnaire respondents were male, below 40 years, and had a pharmacy diploma.
One hundred and three (95%) pharmacies received a simulated client call. Thirteen (13%) calls were determined as probably detected. Detection was less likely if the call was made by one particular caller (perfect prediction), gender of person answering the call was male (adjusted odds ratio, aOR, 0.1, p=0.01); the pharmacy had a publicly available phone number (aOR 0.01, p=0.01), was part of a chain network (perfect prediction), opened 24 hours daily (perfect prediction), or had an on-site laboratory (perfect prediction). Main call outcomes (preventive therapies recommended, options for delivery of recommended products, and referral options) did not vary by detection status (data not shown), therefore all data were included in analysis.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.22.21265188 doi: medRxiv preprint Impact of the pandemic on service providers and pharmacy operations Interview participants described fear and panic when they first heard about COVID-19, especially regarding the risk of exposure to infection: "In the beginning, it sent everybody into a panic mode. Including my entire staff where everybody now thought this is a killer disease and being healthcare workers, they must be people who will be among the victims." -Male pharmaceutical technologist, Nairobi, 105 Some providers were in denial, feeling that the pandemic would never reach "where we are", and others reported that some clients believed the pandemic was "just an avenue maybe for our politicians ready to make some money or something".
The main concerns reported in the questionnaire were: exposure to infection at work (66% of respondents), "keeping my family safe" (22%), meeting financial needs (6%), and being caught up in the curfew (3%).
One hundred (93%) of 108 pharmacies reported pandemic-induced operational challenges (Figure 1). The main supply-side challenge was unavailability and high price of products and the main demand-side challenge was reduced client flow. Supply-side challenges . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.  Interview data corroborated survey findings. Interviewees described temporal changes in client flow, with an initial surge of clients buying vitamins and other immune boosters, followed a few weeks later by a dip in demand to below pre-pandemic levels. The initial surge was accompanied by an increase in self-medication and stockpiling of long-term medications, such as for diabetes. While some of these requests were deemed potentially inappropriate, providers were not always able to effectively intervene: " Demand-side challenges . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted October 26, 2021. A reduction in hygiene-related illnesses was also hailed as a positive public health effect of the COVID-19 preventive measures: "Due to proper maintenance of hygiene, the short-short illnesses have reduced. Washing of hands and stuff."" -Male pharmacist, Nairobi, 103 Adjustments to pharmacy operations Pharmacies adjusted their operations to cope with these challenges and to accommodate the government-instituted prevention measures: 74% reduced opening hours (mean 3 hours [range 1-7]); 38% reduced staffing; and 30% changed their meeting routine ( Table 1). Adjustments did not vary significantly across counties.  Table 2 presents the COVID-19-related products and services that were being provided or requested at pharmacies. Equipment and materials in highest demand were medical masks, N95/ K95 respirators, face shields, hand sanitizer, and thermometers; while the medications in highest demand were vitamin C, multivitamins, and antibiotics. Only 24 (22%) pharmacies reported keeping records related to COVID-19 services. Interview data suggested that many of these products and services were newly introduced following the outbreak, including products like vitamin C and surgical masks that one would have expected to find routinely in a pharmacy.

Demand for COVID-19 testing and provider views towards pharmacy-based testing
Fifty-nine (55% [95% CI 45-65%]) of 108 pharmacies reported ever receiving a client asking for COVID-19 testing, with a median [interquartile range] number of requests in the last week of 2 [1][2][3][4]. At the time of the survey, none of the pharmacies was providing COVID-19 testing; and among the 59 pharmacies experiencing demand for testing, alternatives given to clients included referral to other testing sites (94%) and body temperature measurement (15%). Referral options included: public health facilities (92%), private health facilities (27%), and private labs (13%).
Interviews revealed that test requests were mostly by clients with symptoms, those who suspected they had been exposed, and those who needed certification for travel or employment purposes (eg, drivers and hotel staff).
In interviews, the confidentiality and convenience of the pharmacy setting was reiterated, with some interviewees suggesting home-based self-testing as an option: . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) Among the 42 that were not supportive of pharmacy-based testing, main reasons for opposing were: might be misused (29%), no private space in the pharmacy (24%), risk of infection (15%), protective gear is expensive or not available (12%), there is no demand for the service (10%), and staff are not adequately trained (10%). From the interviews, hesitancy to support pharmacybased testing was mainly related to lack of capacity ("I still need a thorough training…") and fear of infection.

Handling of simulated clients seeking covid-19-related services
When a simulated client called a pharmacy to ask if they have medicines that the client can use to protect themselves from getting COVID-19, 76 (74%) of 103 pharmacies recommended one or more medications: 27% recommended at least one conventional therapy and 71% at least one alternative therapy or nutritional supplement ( Table 3). The main recommendations given to a client suspecting COVID-19 infection were to: get tested, visit the nearest health facility, and selfquarantine at home. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted October 26, 2021. Caller assessment of the interaction . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted October 26, 2021.

Infection prevention during service provision
The main infection prevention measures during service provision were hand hygiene, face masks, disinfection of services and physical distancing ( Table 4). Eighty-five per cent of respondents rated the prevention measures instituted at their pharmacy as either adequate (72%) or very adequate (13%). Only one pharmacy reported that a staff member had ever tested positive for COVID-19, though no pharmacy reported regular testing of staff.

-3 meters between staff and clients through the use of posters and floor/seat markings, serving one client at a time or other crowd control measures, ribbons/ ropes/ glass/ metal-grill/ wooden barriers between client and provider, and home
delivery of products. One pharmacy noted that "patients with cough and related symptoms are always first referred to the health center for covid19 testing before being served further". 2 Majority of pharmacies (95%) were using surgical/medical masks; only 6% reported use of cloth masks. 3 Hand hygiene measures among clients included: handwashing, hand sanitizer, and "no hand-shakes" policy 4 Non-touch payment methods included: mobile money, and credit/ debit cards; one pharmacy reported disinfection of cash. 4 Messages were delivered through: posters (21%), leaflets (5%), television screens (2%), and audio announcements (1%)
Interviewees felt that the official announcement of the first case played was a major source of awareness, though some reported having learnt about the outbreak through radio, television, and social media prior to the first reported case. However, they decried the rapidly changing and sometimes contradictory information, for example "on whether to use or not to use masks".
Fifty-two (48%) pharmacies had at least one staff member trained on COVID-19, with a median [range] of 1 [1][2][3][4][5][6][7][8] providers trained. This did not vary by county or by membership in professional associations. Modalities of training included: on-site in the pharmacy (67% of respondents), virtual (54%), and off-site (13%). Virtual trainings were more prevalent in Kisumu compared to . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.22.21265188 doi: medRxiv preprint Nairobi (76% vs 35%, p=0.04). Training sessions were facilitated by MoH (73%), KPA (51%), PSK (19%), AMREF (9%), and Kenyatta National Hospital (5%). While some interview participants felt they had received "a lot'' of training, others felt the trainings were not very adequate, particularly with regard to inclusion of updated information in follow-up sessions and the fact that most trainings included only the pharmacy in-charge and not the entire staff.

External support and linkages
Interviewees expressed appreciation for guidance from the ministry of health, including the pharmacy and poisons board, who "did some inspection and brought some banners [information materials] on how I can protect myself and the general public." The government-led community awareness-creation activities were also hailed as an important complement to pharmacy efforts in educating the public.
However, examples of direct linkages and material support were rare. One interviewee in Nairobi reported that community health workers from a nearby health centre regularly visited the pharmacy to collect information on number of suspect cases seen. Another interviewee in Kisumu described an instance where an ambulance was sent promptly on reporting a patient in critical condition.
More commonly, a feeling of isolation and neglect prevailed, with pharmacies reporting that they were not receiving much support from the government or other external parties with regard to COVID-19, unlike what happens with other public health programs: "So far being a private facility, we are not receiving any financial support from the organizations in COVID related issues. We don't receive that apart from PS Kenya [NGO supporting TB and HIV programs]." -Male pharmaceutical technologist, Kisumu, 301 Confidence, motivation, and coping mechanisms Majority of questionnaire respondents reported feeling confident (68%) or very confident (19%) about providing COVID-19-related services. Over half felt prepared (51%) or very prepared (7%) about handling a client who presents to the pharmacy with COVID-19 symptoms.
Motivation for supporting delivery of COVID-19 services included: "to help people protect themselves or prevent transmission (75%), "It's part of my job/ I am required to do it" (14%), encouragement and support from professional associations (7%), "to boost the profile of the pharmacy" (2%), and "to make money" (1%). When asked how much they felt their personal role in the COVID-19 response was valued by the community, the majority stated highly valued (68%) or very highly valued (11%) In addition to the operational adjustments described above, interviewees described other coping mechanisms such putting up a shade outside to prevent crowding inside the pharmacy and . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.22.21265188 doi: medRxiv preprint improvisation of hand sanitizers when not available from suppliers. Some interviewees also reported going the extra mile sometimes, eg, by providing free masks to clients.

Provider suggestions and recommendations
Seventy-six (70%) questionnaire respondents highlighted inputs that they felt pharmacies can uniquely contribute to the national response, including: provision of information and counselling to clients (64%), distribution of preventive materials (21%), triage and referral of suspect cases (17%), surveillance of infection rates and community adherence to preventive measures (14%), provision of preventive and curative therapies (12%), and maintenance of long-term therapy (3%). Pharmacies' location within communities was cited as a strength when it comes to surveillance of client behaviour and community awareness creation.
To facilitate these inputs, pharmacies suggested additional support that they would require from the government and other external parties, including: training (34% of respondents), free or subsidized preventive materials such as masks, sanitizers, and thermo guns for own use and distribution to clients (34%), supply management such as price and quality controls (15%), clear guidelines and referral systems (9%), community awareness creation (7%), information materials (6%), financial support such as loan repayment holidays and medical insurance (6%), and testing services (5%).
These suggestions were echoed in interviews. Post-marketing surveillance to ensure quality of products in the market and a streamlined process to capture and transmit information back to the government were pinpointed as key gaps in the pandemic response. Overall, there were earnest calls for more deliberate engagement, oversight, and support supervision by the government and other implementing partners: "If the government could put us, the pharmacists and pharm techs in the you know, treat us like people who are fighting this virus and facilitate this thing it would be a major shift, it would be a major shift." -Male pharmacy in charge, Mombasa, 205

Discussion
In this study among private retail pharmacies (community pharmacies) in Kenya, we describe the impact of the COVID-19 pandemic on operations, COVID-19-related products and services being provided, infection prevention during service prevention, and provider suggestions and recommendations on how to improve pandemic preparedness and response.
The initial weeks after the pandemic reached Kenya were characterized by fear and panic among providers and a surge in client flow. Broadly speaking, the experiences documented in our study can be captured with the apt words of Austin et al in their paper describing the experiences of community pharmacists in Ontario, Canada: "the scale and magnitude of the COVID-19 pandemic . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.22.21265188 doi: medRxiv preprint defied imagination or pre-planning, yet health care professionals like pharmacists were required to continue to provide services and care to patients -and in many cases, expand their repertoire of clinical skills to assume new and even more challenging responsibilities" [15].
The initial surge in demand waned rapidly, with client flow dipping below pre-pandemic levels within a few weeks and remaining depressed until the time of data collection, seven months later. This was surprising as we had hypothesized, before the study, that pharmacies would experience an overall increase in demand as patients redirected their care-seeking from health facilities for fear of exposure to infection and quarantine. Participants attributed the decrease in demand mainly to the movement restrictions instituted by government. However, they also suggested -ironically -that clients, especially those with flu-like symptoms, may have avoided pharmacies for fear of being referred to public facilities and subsequently being quarantined.
Our findings clearly show that community pharmacies in Kenya were undertaking potentially impactful activities to support the response to the COVID-19 pandemic. Despite a feeling of disconnection from the government-led national response, providers expressed great readiness and motivation for provision of COVID-19-related services, similar to providers in Qatar [30], the USA [31], and other settings around the world [32]. The services pharmacies in our study were providing are generally similar to those suggested by experts in the field [22,[33][34][35], as well as those being provided in other low-to-middle income countries [36][37][38][39] and in high-income settings [30,40,41].
However, management of drug shortages and prevention of stockpiling (hoarding) by clients [22,30,33] were reported only as challenges and not as roles that Kenyan pharmacies were playing or could play. Additionally, while 60% of questionnaire respondents said they were receiving queries and providing information and counseling over the phone, other remote services such as online prescribing and home delivery of medications [22,36] were not being widely used or mentioned as potential additional roles. Participants in our study, being frontline workers, may have deemed the carrying out of these innovations as being beyond the sphere of their influence. Similarly, COVID-19 vaccine administration was not mentioned as a potential role, understandably perhaps because a vaccine was not available at the time. But pharmacies in Kenya are actively involved in provision of other vaccines [42,43], hence it is reasonable to hypothesize that COVID-19 vaccines could be delivered through private pharmacies in the future. Interviews with policymakers and captains of industry may be required to probe these topics further.
The finding that no pharmacy was providing COVID-19 testing was unsurprising since no approved point-of-care (POC) tests were available at the time of the study. The more interesting finding is that there was significant demand and support for pharmacy-based testing, similar to what studies in the Middle East [44] and the USA [31] have found. In this context, and given the fact that at least three rapid POC tests have since been approved in Kenya [45], there is an urgent . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.22.21265188 doi: medRxiv preprint need to start developing policies to guide implementation. This will obviate the emergence of an uncoordinated and potentially dangerous service, similar to the HIV testing one that we observed in 2016 before policy guidelines were issued [46,47].
For treatment of COVID-19 symptoms, most pharmacies recommended alternative therapies and nutritional supplements such as vitamin C and zinc, and only about a third suggested conventional therapies with no proven efficacy such as antibiotics [13,14]. This finding is similar to that of a small study in Nairobi in May 2020 which found that despite increased requests for antimalarials and antibiotics, pharmacists recommended alternative therapies [48]. A similar finding was obtained in a multinational survey of Asian pharmacists [35]. While alternative therapies and nutritional supplements are themselves not evidence-based (as no rigorous trials have proven their efficacy against SARS-CoV-2), they have a better safety profile and their useful effects are more biologically plausible since they have been used for ages against the flu virus, itself a coronavirus. These findings are therefore encouraging. However, among Egyptian pharmacies, the antibiotic azithromycin was given to about 40% of presumptive COVID-19 patients with mild to moderate symptoms, albeit based on physician recommendation [49]. Further research in different jurisdictions may be needed to describe and contrast the use of antibiotics and other unproven COVID-19 therapies, in order to inform the promotion of rational drug use and the prevention of antimicrobial resistance.
Our study identified other important gaps in community pharmacies' preparedness and response to the pandemic. Notably, pharmacies did not seem to have any well-defined operational linkages to other units of the healthcare system, including for example, the emergency operations center at the ministry of health. Besides, while only about half of pharmacies had at least one staff member trained on COVID-19, some of those trained felt they had received a lot of training. This suggests a "network effect" that our study was unable to decipher, where a certain group of providers seem to be well-positioned to receive many trainings, while others receive none.
Specific programs will be required to address these gaps and enhance pandemic preparedness. Templates for public-private cooperation in pandemic responses do exist from other settings. Ung et al, for example, report how in one administrative region on the south-east coast of China, pre-existing public-private partnerships between the local government and community pharmacies ensured rapid provision and prevented hoarding of face masks through a government-led "Guaranteed Mask Supply for Macao Residents Scheme" [50]. A qualitative study among pharmacists in Canada, identified key features that predicted community pharmacists' resilience during the early period of the pandemic, including: use of and comfort with technology; early adoption of corporate and professional guidance; emphasis on task-focus rather than multi-tasking; and provision of personal protective equipment [15]. Roll-out of pharmacy-based COVID-19 programs should be accompanied with implementation studies to better uncover aspects of capacity, preparedness, and resilience.
. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021. 10.22.21265188 doi: medRxiv preprint Our study had a number of strengths. First, we took advantage of information collected in a recent study to speed up recruitment and data collection. Second, we used multiple quantitative and qualitative methods to enhance the comprehensiveness and reliability of the findings. Of note, we used simulated clients to objectively assess provider practices, an approach that is gaining acceptance as a gold standard for the measurement of clinical practice quality around the world [51]. One weakness of our study was the low response rate (55%), considering that we targeted pharmacies that had participated in a similar study less than a year previously. Compared to non-participating pharmacies, participating pharmacies had features suggestive of higher operational capacity and interest in public health interventions. While this limits generalizability of our findings, it can also be looked at in positive light. We propose that pharmacies with these characteristics could be identified for inclusion in a "practice and learning network" that could provide a platform for testing implementation strategies for various public health interventions [52-54]. Scale up of such interventions would then target similar pharmacies and include strategies for enhancing operational capacity and interest in public health among the other pharmacies.
In conclusion, we found that private retail pharmacies in Kenya were actively contributing to the COVID-19 response despite an apparent disconnection from the national response program. More deliberate engagement, support and linkages are required. Notably, there is an urgent need to develop policy guidelines for pharmacy-based COVID-19 testing, a service that is clearly needed and which could greatly increase test coverage. Roll-out of this and other pharmacybased COVID-19 programs should be accompanied with implementation research in order to inform current and future pandemic responses.
Research-Nairobi is supported by core funding from the Wellcome Trust [#077092]. This report was published with the permission of Director KEMRI CGMRC.

Patient and Public Involvement
Findings will be shared with the pharmacy sector at the 2021 Pharmaceutical Society of Kenya Annual Conference and a policy brief will be prepared targeting stakeholders in the ministry of health, pharmacy sector and other implementing partners. . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted October 26, 2021. ; https://doi.org/10.1101/2021.10.22.21265188 doi: medRxiv preprint