Study Selection and Study Characteristics

After screening, 25 articles were assessed for eligibility. During the second stage screening process, 12 articles were excluded because they were not peer-reviewed journal articles (ie, news periodicals, letters to the editor, commentary) or did not sufficiently address DTCs. In total, 13 articles satisfied the inclusion and exclusion criteria and were included in this review.^{Reference Lee, Chien-Hung, Pietz and Benecke2-Reference Zerwekh, McKnight and Hupert6,Reference Le, Charney and Gerard11-Reference Gupta, Evans and Heragu13,Reference Weiss, Ngo, Gilbert and Quinn16-Reference Carrico, McKinney and Watson20}

The majority of the articles were descriptive studies (53.85%, n = 7), followed by modeling and simulation studies (30.77%, n = 4), and summary articles (15.38%, n = 2). Five of the descriptive studies were real-world scenarios – where participants were members of the community being served – and 2 of the descriptive studies were hypothetical exercises, where case scenarios were created to mimic/replicate an outbreak. The main aims of the descriptive studies were to assess participant throughput times,^{Reference Zerwekh, McKnight and Hupert6,Reference Le, Charney and Gerard11,Reference Weiss, Ngo, Gilbert and Quinn16,Reference Banks, Crandall and Esquibel17} exposure to toxic carbon monoxide (CO) levels,^{Reference Nicholas, Eileen, Julia and Ron12} clinic usage questions and demographic characteristics of the DTC participants,^{Reference Banks, Vanderjagt and Crandall18} and to evaluate the effectiveness of university organized DTCs in the event of an outbreak.^{Reference Rega, Bork and Chen19} The modeling and simulation studies focused on creating DSSs and decision support tools to be used by decision-makers and public health agencies for preparing and responding to future outbreaks and potential biological attacks.^{Reference Lee, Chien-Hung, Pietz and Benecke2,Reference Ramirez-Nafarrate, Lyon, Fowler and Araz3,Reference Richter and Khan5,Reference Gupta, Evans and Heragu13} The summary articles focused primarily on adverse events associated with DTCs and traditional PODs. Specifically, Carrico et al.^{Reference Carrico, McKinney and Watson20} estimated the probability of an adverse event (ie, syncopial and vehicle accidents) occurring during a DTC, while Rebmann and Coll^{Reference Rebmann and Coll4} outlined infection prevention recommendations for traditional and DTC POD modalities.

Although most articles did not focus primarily on the identification of effective practices and recommendations, each article did address effective practices and recommendations specific to its study. Therefore, extrapolation of effective practices and recommendations was conducted. The effective practices presented in this review refer to practices demonstrated as being effective, but not as a best practice through a formal validation process.²¹ The identified themes and associated number of articles were (1) optimal DTC design and planning via DSSs and decision support tools (5 articles [39%]); (2) clinic layouts, locations, and design aspects (5 articles [39%]); (3) staffing, training, and DTC communication (9 articles [69%]); (4) throughput time (7 articles [54%]); (5) community outreach methods (6 articles [46%]); (6) DTC equipment (6 articles [46%]); (7) infection prevention and personal protective equipment (PPE) (5 articles [39%]); and (8) adverse events prevention and traffic management (7 articles [54%]). Table 1 presents the overall results (ie, themes, effective practices, recommendations, and study findings) from each article.

TABLE 1 Effective Practices, Recommendations, and Study Findings by Article

CO = carbon monoxide; DSS = decision support system; DTC = drive-through clinic; ED = emergency department; EMS = emergency medical services; GA = genetic algorithm; ICS = incident command system; LOS = length of stay; PCP = primary care physician; POD = point of dispensing; PPE = personal protective equipment; PPM = parts per million; RN = registered nurse; USPS = United States Postal Service.

Optimal DTC Design and Planning via DSSs and Decision Support Tools

Five DSSs identified in the literature focused on the DTC design and planning phases, and 1 decision support tool focused on identifying the safest medication for each patient. The DSSs varied by their intended applications, model type (eg, simulation, optimization, combination of simulation and optimization), features (eg, graphical tools, user interface), model complexity, and ease of use. Patient throughput time/speed of dispensing, performance evaluation through large-scale emergency event simulation, and the number of staff needed were the most popular topics for modeling, with every DSS identified evaluating said variables in a way specific to the models’ intended application.^{Reference Lee, Chien-Hung, Pietz and Benecke2,Reference Ramirez-Nafarrate, Lyon, Fowler and Araz3,Reference Richter and Khan5,Reference Zerwekh, McKnight and Hupert6,Reference Gupta, Evans and Heragu13}

Specifically, Lee et al.^{Reference Lee, Chien-Hung, Pietz and Benecke2} developed a DSS to design and simulate various DTC layouts to optimally allocate staff and resources based on a specified throughput time. Gupta et al.^{Reference Gupta, Evans and Heragu13} combined generalized discrete-event simulation and optimization to estimate the average patient throughput time based on the expected number of vehicles, the number and length of consent form and vaccination lanes, and the number of staff per lane. Ramirez-Nafarrate et al.^{Reference Ramirez-Nafarrate, Lyon, Fowler and Araz3} proposed a flexible algorithm to select optimal PODs from a candidate list and allocate the optimal number of staff to each dispensing station so the average travel and waiting times are minimized. Zerwekh et al.^{Reference Zerwekh, McKnight and Hupert6} used the Bioterrorism and Epidemic Outbreak Response Model (BERM), which predicts the number and type of staff needed while providing feedback on patient throughput and queuing times. Last, Richter and Khan^{Reference Richter and Khan5} applied a qualitative value additive model with an objective hierarchy to assess the trade-offs between using various dispensing modalities (eg, DTCs, traditional walk-in clinics, pharmacies, and USPS delivery) for a specific population by determining their overall effectiveness based on POD dispensing speed, staff requirements, and security needs. In addition to the applications mentioned previously, unique applications and features that distinguish each DSS and decision support tool can be found in Table 2.

TABLE 2 Decision Support Systems and Decision Support Tools

DTC = drive-through clinic; GA = genetic algorithm; GIS = graphical information system; POD = point of dispensing; SEIR = susceptible, exposed, infectious, and recovered; SEPAIR = susceptible, exposed, infectious, asymptomatic, symptomatic, and recovered; USPS = United States Postal Service.

Clinic Layouts, Locations, and Design Aspects

Various clinic layouts and locations were ascertained from the descriptive studies. They varied by the number of stations, types of stations, number of processing lanes, and clinic locations based on specific target populations and goals. The types of stations identified included (1) triage, (2) registration, (3) screening, (4) evaluation/examination, (5) special needs/services, (6) dispensing/vaccination, (7) emergency medical services/post-vaccination observation, (8) discharge, and (9) medical records. Locations included a large covered parking-structure,^{Reference Weiss, Ngo, Gilbert and Quinn16} open parking lots,^{Reference Le, Charney and Gerard11,Reference Banks, Crandall and Esquibel17,Reference Rega, Bork and Chen19} a large stadium,^{Reference Zerwekh, McKnight and Hupert6} and an enclosed school bus garage.^{Reference Nicholas, Eileen, Julia and Ron12}

Beneficial clinic design aspects were identified. A large spatial arrangement was found to be beneficial due to the large distance between stations, which allowed vehicles to stack up.^{Reference Zerwekh, McKnight and Hupert6} The stacking of vehicles allowed the staff to process bottlenecked vehicles more rapidly, therefore, increasing throughput.^{Reference Zerwekh, McKnight and Hupert6} It was recommended to locate DTCs near major intersections and streets to increase visibility and accessibility.^{Reference Le, Charney and Gerard11} Providing a screening/triage station at the beginning of the DTC was found to be beneficial because it allowed staff to identify patients in need of special assistance or who needed to be diverted and transported to a hospital.^{Reference Rebmann and Coll4,Reference Zerwekh, McKnight and Hupert6,Reference Weiss, Ngo, Gilbert and Quinn16} Critically ill patients could be transported quickly to a hospital by exiting the normal processing lanes via an emergency bypass lane.^{Reference Zerwekh, McKnight and Hupert6,Reference Weiss, Ngo, Gilbert and Quinn16} An evaluation station located between registration and dispensing stations was found to be beneficial in determining the correct medication for each patient.^{Reference Zerwekh, McKnight and Hupert6} Using various colors for each tent station helped patients identify specific stations.^{Reference Zerwekh, McKnight and Hupert6}

Staffing, Training, and DTC Communication

Determination of staff needed for DTCs, proper staff training, and establishment of communication within DTCs were crucial planning and implementation steps identified in the literature. Utilization of the aforementioned DSSs and decision support tools (ie, RealOpt, BERM, Gupta et al.’s DSS, Ramirez-Nafarrate et al.’s DSS, and Richter and Khan’s DSS) was found to be an effective practice in determining the optimal number of staff needed for each DTC and station within each DTC.^{Reference Lee, Chien-Hung, Pietz and Benecke2,Reference Ramirez-Nafarrate, Lyon, Fowler and Araz3,Reference Richter and Khan5,Reference Zerwekh, McKnight and Hupert6,Reference Gupta, Evans and Heragu13} Allocating registered nurses to clinical stations where they were more familiar with clinical terms (ie, compared to nonclinical staff) was also found to be effective.^{Reference Zerwekh, McKnight and Hupert6}

Effective training strategies and recommendations were mentioned in the literature. They included multiple trainings on SNS preparedness plans, leadership, safety, security, clinic design and function, staff roles, staff responsibilities, mental health support issues, event times, DTC communication and radio etiquette, and infection prevention strategies.^{Reference Rebmann and Coll4,Reference Zerwekh, McKnight and Hupert6,Reference Rega, Bork and Chen19} Training on implementing infection prevention and occupation health strategies, such as exposure prevention, handling and disposing sharps, using cold-chain techniques (ie, refrigerate vaccines), proper screening, hand hygiene techniques, environmental decontamination, selecting and using the correct PPE, respiratory etiquette, and spatial distancing techniques, were recommended to decrease disease transmission.^{Reference Rebmann and Coll4} Preregistration of the staff, staff’s families, and first responders was found to be beneficial in training staff on efficiently administering registration forms.^{Reference Zerwekh, McKnight and Hupert6}

DTC communication practices identified were using 2-way radios,^{Reference Zerwekh, McKnight and Hupert6,Reference Weiss, Ngo, Gilbert and Quinn16,Reference Rega, Bork and Chen19} inviting HAM radio operators to assist in communications,^{Reference Rega, Bork and Chen19} communicating and reporting through an Incident Command Communication Center,^{Reference Rebmann and Coll4,Reference Rega, Bork and Chen19} and establishing a Joint Information Center for communication between multiple stakeholders and PODs.^{Reference Rega, Bork and Chen19}

Throughput Time

Increasing overall participant throughput while decreasing participant length of stay was identified as a critical effective practice. Staffing strategies included optimal allocation of staff via utilization of the aforementioned DSSs and decision support tools,^{Reference Lee, Chien-Hung, Pietz and Benecke2,Reference Ramirez-Nafarrate, Lyon, Fowler and Araz3,Reference Richter and Khan5,Reference Zerwekh, McKnight and Hupert6,Reference Gupta, Evans and Heragu13} staff increases when and where necessary,^{Reference Zerwekh, McKnight and Hupert6} and thorough staff training on the registration form format (ie, registration station was often found to be the most time-consuming station).^{Reference Zerwekh, McKnight and Hupert6} Moreover, registration forms provided in large, single-sided print,^{Reference Zerwekh, McKnight and Hupert6} verbally administered surveys/registration forms,^{Reference Zerwekh, McKnight and Hupert6,Reference Le, Charney and Gerard11} and forms that were completed while patients were in queue helped decrease patient throughput times.^{Reference Rega, Bork and Chen19}

Other effective strategies and recommendations identified were vehicle stacking at each station, which allowed the evaluation of multiple vehicles simultaneously,^{Reference Zerwekh, McKnight and Hupert6,Reference Banks, Crandall and Esquibel17} specification of the optimal number of patients per vehicle (ie, 3 to 4 patients per vehicle based on resources and DTC capacity) and encouragement of carpooling,^{Reference Zerwekh, McKnight and Hupert6,Reference Banks, Crandall and Esquibel17} small trays with supplies carried by multiple staff members allowed vaccination of multiple patients per vehicle,^{Reference Banks, Crandall and Esquibel17} and having plans that address inquisitive patients in a way that decreases questions and maximizes throughput.^{Reference Zerwekh, McKnight and Hupert6} Last, multiple PODs (ie, combination of traditional walk-ins and DTCs) across a region could also decrease throughput time.^{Reference Zerwekh, McKnight and Hupert6}

Community Outreach Methods

To inform and attract the community, a diverse advertising campaign that used multiple outlets was found to be effective. The outlets used and recommended were newspapers,^{Reference Zerwekh, McKnight and Hupert6,Reference Le, Charney and Gerard11,Reference Banks, Vanderjagt and Crandall18} radio stations,^{Reference Zerwekh, McKnight and Hupert6,Reference Le, Charney and Gerard11,Reference Weiss, Ngo, Gilbert and Quinn16,Reference Banks, Vanderjagt and Crandall18} television ads,^{Reference Banks, Vanderjagt and Crandall18} websites,^{Reference Le, Charney and Gerard11,Reference Banks, Vanderjagt and Crandall18} student/staff e-mail,^{Reference Banks, Vanderjagt and Crandall18} flyers,^{Reference Zerwekh, McKnight and Hupert6} and ads posted at primary care clinics.^{Reference Le, Charney and Gerard11} Banks et al.^{Reference Banks, Vanderjagt and Crandall18} surveyed DTC participants on how they were informed of the DTC to ascertain the most effective advertising modalities. Newspaper and television ads were discovered to be most effective, with 33.85% and 33.21% of participants reporting they heard about the DTC through these 2 modalities, respectively.^{Reference Banks, Vanderjagt and Crandall18} These advertising modalities were followed by word of mouth (19.6%), university website (8.17%), radio (2.81%), and other websites (2.36%). Rega et al.^{Reference Rega, Bork and Chen19} noted the importance of confirming accurate and consistent DTC opening and closing times posted by various media outlets. Furthermore, incentives such as free vaccinations^{Reference Banks, Vanderjagt and Crandall18} and child car seat fittings^{Reference Le, Charney and Gerard11} were used to increase community participation.

In addition to informing the public about DTCs, local radio stations dedicated to public safety broadcasts can routinely update the community on projected wait times, DTC instructions, and live reports.^{Reference Weiss, Ngo, Gilbert and Quinn16} Informative/educational brochures should be available to inform the public about the DTC event, the specific disease outbreak, infection prevention strategies, strategies to identify infected individuals, methods to limit exposures and high-risk populations, and when and where someone should seek prophylaxis.^{Reference Rebmann and Coll4,Reference Zerwekh, McKnight and Hupert6,Reference Weiss, Ngo, Gilbert and Quinn16}

Banks et al.^{Reference Banks, Vanderjagt and Crandall18} recommended using surveys in future, non-emergency DTCs to ascertain participating community member demographic characteristics so members of the community who do not participate could be identified and better targeted. Furthermore, it is recommended that the surveys be culturally sensitive (ie, offered in various languages using culturally appropriate terminology) to increase the accuracy of responses and should incorporate other critical variables associated with participation and vulnerable populations, such as insurance coverage and high-risk medical conditions.^{Reference Banks, Vanderjagt and Crandall18}

DTC Equipment

DTC equipment and supplies used and recommended in the articles included general equipment,^{Reference Rebmann and Coll4,Reference Zerwekh, McKnight and Hupert6,Reference Nicholas, Eileen, Julia and Ron12,Reference Weiss, Ngo, Gilbert and Quinn16,Reference Banks, Crandall and Esquibel17,Reference Rega, Bork and Chen19} emergency equipment,^{Reference Weiss, Ngo, Gilbert and Quinn16} traffic control supplies,^{Reference Zerwekh, McKnight and Hupert6,Reference Weiss, Ngo, Gilbert and Quinn16,Reference Rega, Bork and Chen19} prophylaxis and medication,^{Reference Rebmann and Coll4,Reference Zerwekh, McKnight and Hupert6,Reference Weiss, Ngo, Gilbert and Quinn16,Reference Banks, Crandall and Esquibel17} documentation/forms,^{Reference Rebmann and Coll4,Reference Zerwekh, McKnight and Hupert6,Reference Nicholas, Eileen, Julia and Ron12,Reference Weiss, Ngo, Gilbert and Quinn16-Reference Banks, Vanderjagt and Crandall18} environmental decontamination supplies and PPE,^{Reference Rebmann and Coll4,Reference Zerwekh, McKnight and Hupert6,Reference Nicholas, Eileen, Julia and Ron12,Reference Weiss, Ngo, Gilbert and Quinn16} and patient care supplies.^{Reference Weiss, Ngo, Gilbert and Quinn16} Public health agencies and decision-makers should assess the equipment needed for their specific DTC based on the event (ie, disease outbreak, biological attack, seasonal flu vaccination, natural disaster), target population, location and size of the DTC, capacity restrictions, and resources available. Table 3 presents the specific DTC and PPE identified or recommended in each article.

TABLE 3 DTC Equipment and Personal Protective Equipment Identified or Recommended

DTC = drive-through clinic; ED = emergency department; HEPA = high efficiency particulate air; MDI = metered-dose inhaler.

Infection Prevention and PPE

A unique advantage of DTCs over traditional walk-in clinics is the social distancing of patients (ie, vehicle acts as an isolation chamber), which mitigates providers’ exposure to infected patients.^{Reference Rebmann and Coll4,Reference Zerwekh, McKnight and Hupert6,Reference Nicholas, Eileen, Julia and Ron12,Reference Gupta, Evans and Heragu13} Isolating patients helps reduce infection propagation, especially at PODs where large numbers of infected patients may be in close proximity. Other strategies and recommendations identified in the literature to prevent infection propagation were screening and triage (ie, patients and staff),^{Reference Lee, Chien-Hung, Pietz and Benecke2-Reference Rebmann and Coll4,Reference Zerwekh, McKnight and Hupert6,Reference Weiss, Ngo, Gilbert and Quinn16} infection prevention training,^{Reference Rebmann and Coll4,Reference Weiss, Ngo, Gilbert and Quinn16,Reference Rega, Bork and Chen19} proper hand hygiene,^{Reference Rebmann and Coll4} occupational health techniques,^{Reference Rebmann and Coll4} sufficient PPE provision,^{Reference Rebmann and Coll4,Reference Zerwekh, McKnight and Hupert6,Reference Weiss, Ngo, Gilbert and Quinn16} and potential disease-propagation evaluation (eg, RealOpt) within the DTC pre-event so mitigating strategies could be employed.^{Reference Lee, Chien-Hung, Pietz and Benecke2}

One study recommended screening and triage be conducted at the beginning of the DTC to divert critically ill patients to a hospital to receive specialized care.^{Reference Weiss, Ngo, Gilbert and Quinn16} Additionally, screening of the clinic staff should be conducted periodically. If a staff member presents symptoms of infection, prompt removal from the DTC is recommended.^{Reference Rebmann and Coll4} Multiple screening strategies were recommended, including visual screening, measuring temperature, direct questioning, or a combination of these techniques.^{Reference Rebmann and Coll4}

Proper hand hygiene, environmental decontamination, and providing PPE were critical aspects of infection prevention in DTCs. To decontaminate hands, physically washing with soap and utilization of alcohol-based sanitizers are effective.^{Reference Rebmann and Coll4} Environmental decontamination should involve Environmental Protection Agency registered disinfectants or a 0.5% bleach solution.^{Reference Rebmann and Coll4} Last, PPE should be appropriate to the task performed and the transmission route of the disease/pathogen.

Adverse Events Prevention and Traffic Management

In addition to infection prevention, adverse event prevention was identified as crucial to DTC implementation. Possible adverse events included CO exposure,^{Reference Nicholas, Eileen, Julia and Ron12,Reference Weiss, Ngo, Gilbert and Quinn16} vehicle accidents,^{Reference Carrico, McKinney and Watson20} syncopal episodes,^{Reference Carrico, McKinney and Watson20} adverse reactions to medications,^{Reference Rebmann and Coll4,Reference Zerwekh, McKnight and Hupert6} aggressive pet interactions,^{Reference Rega, Bork and Chen19} and other issues such as lane blockage and delays in the transport of critically ill patients.^{Reference Rebmann and Coll4,Reference Weiss, Ngo, Gilbert and Quinn16,Reference Rega, Bork and Chen19} The likelihood of these adverse events occurring depended on the specific DTC location (ie, indoor versus outdoor) and the target population served.

Exposure to toxic levels of CO was a concern primarily for indoor/sheltered DTCs that lacked sufficient ventilation. Weiss el at.^{Reference Weiss, Ngo, Gilbert and Quinn16} instructed participants to shut off their vehicle before staff approached. Nicholas et al.^{Reference Nicholas, Eileen, Julia and Ron12} recommended the identification of vehicles in disrepair with the potential to emit high levels of CO before entrance into the indoor DTC and to process these vehicles outside or in an expedited fashion. Moreover, it was suggested to purchase CO monitors or to collaborate with local agencies (ie, fire departments, health departments) that can provide CO monitors to be worn by staff.^{Reference Nicholas, Eileen, Julia and Ron12}

Contraindications were avoided through utilization of a medication algorithm tailored to the specific medications intended to treat the disease/agent that were addressed by the DTC. For example, Zerwekh et al.^{Reference Zerwekh, McKnight and Hupert6} used a medication algorithm to select the safest medication (ie, doxycycline or ciprofloxacin) for each patient based on the patient’s medical history. Using this algorithm, 98.9% of dispensed medications were the safest option.^{Reference Zerwekh, McKnight and Hupert6}

Syncopal episodes following vaccinations have been documented.^{Reference Carrico, McKinney and Watson20} Although rare, consideration of these events is necessary for implementing DTCs.^{Reference Carrico, McKinney and Watson20} To avoid these events and/or to mitigate their impacts, it was recommended to provide an observation station following prophylaxis to monitor patients.^{Reference Weiss, Ngo, Gilbert and Quinn16,Reference Rega, Bork and Chen19,Reference Carrico, McKinney and Watson20} Further, safety monitoring teams could be dispatched to roam around the DTC to monitor vaccinated patients.^{Reference Rega, Bork and Chen19}

Transport of critically ill patients to a clinic or hospital to receive specialized care was an effective practice. If the DTC was located adjacent to a hospital, providing on-site transportation, such as golf carts equipped with stretchers, was found to be effective.^{Reference Weiss, Ngo, Gilbert and Quinn16} If the DTC was not located adjacent to a hospital, an emergency exit/bypass lane should be provided for vehicle exit and emergency vehicle access.^{Reference Zerwekh, McKnight and Hupert6,Reference Rega, Bork and Chen19}

Additional adverse event prevention strategies identified were vaccination of patients outside vehicles if patients were accompanied by aggressive pets,^{Reference Rega, Bork and Chen19} tents anchored to the ground,^{Reference Zerwekh, McKnight and Hupert6} large signs used to direct special needs patients,^{Reference Zerwekh, McKnight and Hupert6} provision of mental health resources,^{Reference Rega, Bork and Chen19} and the presence of security to protect medications and vaccinations.^{Reference Zerwekh, McKnight and Hupert6}

Last, traffic management was a critical component of effective DTC implementation. To best address this issue, the aforementioned DSSs were used to optimally design and select DTC layouts and locations. Additional traffic management strategies included the presence of police/security to direct traffic,^{Reference Zerwekh, McKnight and Hupert6,Reference Weiss, Ngo, Gilbert and Quinn16,Reference Rega, Bork and Chen19} designated lanes and stopping points with clear and visible signs,^{Reference Zerwekh, McKnight and Hupert6,Reference Weiss, Ngo, Gilbert and Quinn16} identification of stations and locations through the use of various colored balloons,^{Reference Rega, Bork and Chen19} traffic flow management to prevent the backup of vehicles,^{Reference Weiss, Ngo, Gilbert and Quinn16} pre-establishment of towing procedures,^{Reference Rega, Bork and Chen19} designation of a space for special needs vehicle processing,^{Reference Zerwekh, McKnight and Hupert6,Reference Banks, Crandall and Esquibel17} the use of barricades/cones to direct traffic,^{Reference Zerwekh, McKnight and Hupert6,Reference Weiss, Ngo, Gilbert and Quinn16} and the provision of sufficient space to allow staff to move freely.^{Reference Zerwekh, McKnight and Hupert6}