Optimizing the Mix of Dedicated and Non-Dedicated Paratransit Services

Ned Einstein
Transportation Alternatives
New York, NY

One important dynamic of paratransit system design still largely misunderstood is that of the relationship between dedicated and non-dedicated service. An exaggerated reflection of this misunderstanding is illustrated by the considerable number of complementary paratransit service (CPS) programs which provide service on only one of these bases.

Dedicated vehicles provide service only to clients of a specific program or funding agency (e.g., a transit agency, under the ADA), or a group of such programs or agencies (e.g., under a coordination or consolidation arrangement), for specific blocks of time (which may be adjusted daily in some operating structures). In contrast, non-dedicated vehicles effectively fade in and out of service to these clients, irregularly or methodically alternating or integrating its provision with service to non-agency-affiliated clients, including the general public. The two most common examples of non-dedicated services are:

  • Taxi services alternating or integrating service to special clients with service to general public riders.
  • Non-emergency medical transportation (NEMT) services integrating CPS with NEMT clients traveling to or from the same common destinations (hospitals, clinics, etc.), and/or in some cases, transporting the CPS clients to other destinations.

The institutional characteristics of even these simple models illustrate the constraints which these modes’ inclusion in CPS systems may create:

  • In most taxi-based services, passengers travel on an exclusive-ride basis, and taxi meters commonly used to calculate their fares can only do so accurately for a single client at a time. The introduction of CPS clients with exclusive-ride taxi riders upsets this neat model for the latter, since their fares have been established for exclusive rides, not for traveling with strangers. Similarly, taxi regulations have guaranteed them directness-of-routing – at least in theory. When trips for both types of clients are integrated, many fare structures (particularly meter-based fares) cannot equitably apportion trip costs among them. Yet even where zonal structure render this latter concern moot, the fare structure cannot eliminate the issue of increased travel times. To avoid these problems, taxis generally provide only exclusive-ride service to CPS clients. However, this orientation conflicts with the shared-ride nature of CPS, and may result in the actual or perceived provision of a higher level of service to selected clients – depending on the system’s rationale for mixing dedicated and non-dedicated services, and assigning passengers to each of them.
  • NEMT services, in contrast, are already shared-ride-oriented, and the integration of CPS with NEMT clients is often seamless at the operating level, since most CPS clients transported on NEMT services travel to and from the same destinations. However, many individuals within both sets of clients are often eligible for both programs, and integrating them exposes both these clients and their funding agencies to increased opportunities for mode splitting them from one program to another – a competition which can generate labyrinthine administrative efforts, and which may be disruptive to the clients bounced from one funding program to another, and forced to comply with separate and different procedures for reserving essentially the same trips provided by the same drivers on the same vehicles (not to mention the confusion associated with CPS transportation provided to NEMT clients on different vehicles).

Compounding such difficulties is the fact that both taxicab and NEMT providers are generally reimbursed, at the vehicle, driver and trip levels, on a unit cost basis – typically derived as $X per trip plus $Y per mile. Even when CPS agencies compensate these providers similarly, the thickened densities these additional clients realize unavoidably creates cross-subsidies, since the rates were derived to cover the costs of providing trips to only the non-CPS clients. Exacerbating this cross-subsidy, CPS agencies often pay NEMT and taxi providers lower, discounted rates for these insertions, since the operating costs are already fully-paid for, and the additional trips are largely “gravy” (exceptions involve marginally-increased fuel and maintenance costs). In real life, while CPS agencies are obviously aware of these dynamics, they may not be disclosed or acknowledged to the regulatory agencies or commissions responsible for establishing the modes’ rate structures.

Non-Dedicated Service Roles and Goals

Regardless of the modes used to provide non-dedicated service, a number of approaches have been employed to assign them to specific roles in the overall service structure, depending on a mix of planning, operating, institutional, geographic, political and/or other goals and constraints. Some of the most common of these roles and/or goals, genuine or illusory, have included:

  • To lower costs in low-density parts of the service area
  • To lower costs during low-density periods of demand
  • To simplify the trip negotiation and scheduling processes
  • To reduce the dedicated vehicle spare ratio
  • To reduce the cadre of back-up drivers
  • To ward off or resolve class action lawsuits related to trip denials
  • To ward off or resolve class action lawsuits related to on-time performance deficiencies
  • To respond to institutional constraints and political pressure
  • To improve overall system on-time performance
  • To accelerate implementation (pre-July, 1997)
  • To rationalize dedicated vehicle shift assignments
  • To minimize overtime for dedicated vehicle drivers
  • To minimize liability (an illusion, since the “Stands-in-the-Shoes” doctrine of the ADA trumps any notions of lead agency insulation via their engagement of “independent contractors”)
  • To satisfy the legitimate and whimsical needs and preferences of certain client sub-groups (e.g., ambulatory elderly individuals can board and alight from taxi sedans more easily than vans)
  • To provide isolated trips difficult or impractical to group with most others
  • To handle “overflow” or otherwise accommodate unusual and/or unpredictable fluctuations in demand and/or shortages of supply (e.g., driver absenteeism aberrations, such as those following three-day week-ends or during “flu season”)
  • To accommodate dedicated vehicle maintenance and inspection needs
  • To “pad” periods during which dedicated fleet expansions are pending (i.e., vehicles on order)
  • To “assist” dedicated vehicles behind schedule (by TXing or dispatching trips away from them)
  • To fetch stranded passengers (including those legitimately or illegitimately identified as “no-shows”)
  • To transport problem clients (e.g., autistic children and adults) whose safety, and that of fellow-passengers, is jeopardized by their travel on shared-ride services

While most of these roles and goals and their overlapping permutations are legitimate, focusing on them is not only dysfunctional but, frankly, child-like in its sophistication. The single approach which encompasses all the legitimate roles and goals incidentally, and which is far more important to performance and cost, is to deploy non-dedicated vehicles to optimize the performance of the systems’ dedicated vehicles. The basic mechanism for achieving this goal is to simply sort those trips into non-dedicated vehicles which are incompatible with those trips efficiently provided by dedicated vehicles, and which would otherwise have to be “shoe-horned” into dedicated vehicle schedules. Realizing the optimum mix under such an approach simply involves sorting trip requests and the resulting trips into those modes which collectively result in the lowest operating costs – costs which can obviously, easily and continually be measured.

Factors Related to the Optimum Mix

The optimum the mix of dedicated and non-dedicated service reflects a myriad of issues and concerns, as well as both system and service area characteristics. Despite all these considerations, however, a strong argument can be made that there is no need to establish the optimum mix as an operating matter: With a crude, intuitive starting point, and an acknowledgement that the optimum mix is a notion that changes from day to day, if not from hour to hour, anyway – the optimum mix du jour – at least at the outset of any given operating day – can be reached through practical experience in a matter of days. Of course, for this to occur, system decision-makers must understand the variables which govern system performance, and system management must be free to adjust the system accordingly.

All this is true largely because the essence of non-dedicated service is that – if one will excuse the redundancy – it is non-dedicated. Of course, the levels of dedicated service cannot always be as easily adjusted, particularly in the short run. Apart from obvious differences in performance from system to system, and differences in operating variables like shift assignment and maintenance, dedicated vehicle service is largely a function of the size of the driver pool and the fleet. Notwithstanding large and flexible cadres of back-up drivers and spares awaiting assignment and deployment, respectively, the dedicated vehicle component of service obviously cannot be adjusted as quickly as the non-dedicated component. Therefore, the need to establish a crude mix of dedicated and non-dedicated service as a planning matter is a legitimate concern.

As a practical matter, the need for a reasonable starting point for creating the optimum mix also exists as a budgetary matter: While the ADA requires the provision of every trip, transportation budgetary processes are rarely this flexible or open-ended, apart from responses to emergencies. Requesting a sizeable “contingency fee” in the budget could place transit agency senior management in an awkward position – particularly if and when dramatic increases in demand emerged during the period to which the budget applied. This likely scenario must also factor in the primitive levels of understanding which board members and elected officials typically have about the dynamics of paratransit operations and performance – dynamics which have eluded many transit agency senior management and staff.

U.S. transit agencies are particularly vulnerable to sudden fluctuations in demand because of the approach to cost control most of them employ: Instead of controlling costs by optimizing performance, most ADA-era transit agencies have controlled, and continue to control, costs by limiting access to service. While the clever employment of ride-limiting barriers to undercut the Spirit and Letter of the ADA are just beginning to erode through the filing of, and progress in, class action lawsuits, the fluctuations in demand resulting from the suppressed satisfaction of demand can translate into dramatic and unpredictable explosions of cost: In response to recent complaints about its alleged six percent trip denial rate, the New York City Transit Authority modified its system to yield a zero denial rate. Instead of triggering the six percent increase in demand which the arithmetic might otherwise suggest – at least to those not terribly savvy about ADA compliance efforts – demand increased by 300 percent. Because the NYCTA relies exclusively on dedicated vehicles – despite one borough possessing one of the planet’s largest taxicab fleets – this increased demand could not be addressed as quickly as it was articulated. To be fair, addressing such dramatic increases in demand even by deploying non-dedicated vehicles would have proven problematic: Over its many protests, the Manhattan taxi industry was only recently ordered to increase its fleet by 900 vehicles. More problematic, of course, is the fact that the radio-less Manhattan taxi fleet only “cruises” (i.e., does not accept telephone, facsimile or e-mail trip requests, and dispatchers have no means of conveying them), and cabbies are notorious for screening out or otherwise avoiding passengers who appear less likely to tip. Such practical realities illustrate the range of concerns that must be factored into the decision to even involve non-dedicated vehicles, much less establish their levels of participation within an optimum mix.

Apart from both the roles and goals of dedicated and non-dedicated services, and the geographical, socio-economic, operating and institutional characteristics of a system and its service area, defining the optimum mix of these services mirrors the same six principles which govern the performance of all forms and types of demand-responsive service:

  • Ratio of fleet to user density
  • Pre-scheduling format
  • Eligibility
  • Trip purpose
  • Service concept
  • Rates and costs

When one examines the characteristics of dedicated versus non-dedicated service with respect to these variables, it is easy to understand how dramatically the mix of both modes – much less their inclusion in the service structure altogether – affects system performance:

  • Ratio of Fleet to User Density. Most urban service areas, and a large number of suburban ones, contain far more non-dedicated than dedicated vehicles irrespective of their involvement in the provision of service to program-affiliated clients. As a result, the inclusion of non-dedicated vehicles greatly thickens fleet density, shortening overall deadhead time and mileage, improving on-time reliability (to the degree that this aspect of service is monitored, evaluated, supervised and enforced), and, at least indirectly, reducing travel time (since service to smaller loads is more practical and affordable).
  • Pre-Scheduling Format. Many U.S. paratransit systems exclude the provision of any immediate-response service – largely because doing so comprises a formidable and multi-dimensional ride-limiting barrier. Far more elderly ambulatory clients would use the system if most trips were provided by taxicab – especially since most trips would likely be provided on an exclusive-ride basis. Unfortunately, the downsides of such an approach include the inability to fill in gaps created by late cancellations and “no-shows,” the tendency for many passengers to schedule trips which they merely might need (but may not need when the day of the trip arrives), the operating chaos created when dedicated vehicles must be TXed to “cover” those behind schedule or broken down, and the vulnerability to class action lawsuits like Richman v. SEPTA and Anderson v. Rochester-Genesee Regional Transportation Authority (and the risk that a que tam motion could be filed against the transit agency) based on trip denials – accommodations for which the dedicated fleet would have to be equipped and staffed to levels beyond those fluctuations even predictable. More importantly, as noted elsewhere in this document, the need to shoehorn every last trip into dedicated vehicle service has profound implications for operations, performance, safety and liability.
  • Eligibility. Particularly in service areas containing disproportionate numbers of ambulatory elderly riders, the inclusion of non-dedicated service understandably translates into increased demand – although one suspects that many deterrents to their ridership will eventually be eroded by class action lawsuits. This is particularly true where the return trip segments cannot be pre-scheduled – but must be accommodated in order to comply with the ADA’s prohibition of (at least) overt trip denials. At the same time, a large fleet of taxicabs operating in a dense urban area can provide far more trips of certain types (e.g., grocery shopping) to certain subgroups (e.g., frail elderly ambulatory individuals) than can dedicated vehicles for the same costs.
  • Trip Purpose. Trip purpose operates as a performance determinant largely because different types of trips can be provided more efficiently according to subscription versus pre-scheduled demand-responsive versus immediate-response formats, and because trips of different types possess different characteristics and dynamics. As a salient example, most individuals’ grocery shopping needs can be med by the nearest (or at least a reasonably proximate) supermarket, whereas they cannot receive medical care from the nearest doctor. (This dynamic makes medical trips longer, more difficult to group, and many of them better candidates for provision by non-dedicated vehicles.) Similarly, it is easier to predict and thus pre-schedule a shopping trip than a doctor’s visit related to a sudden illness. (Under the ADA, clients can of course make individual shopping trips for individual items. But there is nothing preventing lead agencies from charging lower fares for alternate types of service, much less subscription service, as the ADA Paratransit Handbook points out (see p. V-2).
  • Service Concept. Because the arrangement of the provision of service in time and space comprises a major determinant of performance, both the inclusion of dedicated and non-dedicated services, and their mix within this inclusion, unavoidably affect this variable.
  • Rates and Costs. The least-costly type of service for any given trip depends largely on both the dimensions of the trip (i.e., its origin, destination and target pickup or drop-off times) and those of other trips (particularly these immediately surrounding it). Because deadhead costs are built into the rates of taxi and NEMT providers, for example, their costs for providing out-of-the-way trips may be dramatically lower than those of dedicated vehicles for the same trips – even where the costs for passenger time or mileage are higher.

In simple terms, optimizing the mix of dedicated and non-dedicated service involves a determination of which of these two forms of service can provide each trip such that overall trip costs are lowest – and assigning each trip to the mode which will result in the lowest total costs. Contrary to a common cliché, doing this is actually easier than it sounds. This is because, again, one of these forms is non-dedicated, and as such, can not only be assigned to provide a trip at the last minute, but can be assigned to provide any trip that is reasonably or unreasonably difficult to provide on a dedicated basis.

Along these lines it is also important to recognize that the provision of virtually all taxi-based service occurs on an immediate-response basis: Pre-scheduled trips are simply retained (in “slots” or physical compartments in analog systems, and as “files” in digital systems), and dispatched into the nearest vehicle shortly before their scheduled pickup times. In simple terms, the provision of pre-scheduled trips (much less subscription service) by taxi-based providers is not only pointless as a performance matter, but an annoyance as an operating matter – particularly since most service is dispatched rather than scheduled to begin with, and since taxi dispatchers do not conceive of or articulate operating functions in the same terms or ways as paratransit system schedulers. Ignoring such realities has led to much abuse, not to mention occasional class action lawsuits: A principle cause of the performance-related problems underlying Flores v. LACMTA was at least one taxi provider’s practice of ignoring pre-scheduled trip requests altogether – and simply dispatching a vehicle to fetch the stranded client when he or she telephoned to inquire about the missed pickup.

Optimal Starting Points

As noted, the starting point for developing an optimal mix of dedicated versus non-dedicated service is largely system-specific. However, also as noted, this caveat is hardly problematic. This is because one can easily begin with more non-dedicated service than the optimum, and simply peel its levels back as schedulers are able to provide more trips with dedicated vehicles and thicken the demand density of this service component. As such, the mix changes through continued refinements in the trip reservation and negotiation, scheduling and dispatching processes and, to a lesser degree, through increased driver familiarity with the service area, and from management activities such as monitoring, evaluation, supervision and training.

Such an approach is not without its problems. Many CPS clients prefer non-dedicated service (since it often involves exclusive rides), and may have to be “weaned” from it if and when they are mode-split into dedicated services. This weaning comprises a practical problem which may have political overtones (e.g., clients complaining to their city council representatives). Yet starting new service modifications with larger-than-necessary complements of non-dedicated service also has its strong points: A large fleet, whose drivers are generally familiar with the service area, can provide the full complement of additional trips needed almost immediately – providing time to phase in more dedicated service and refine its performance. Such characteristics of non-dedicated service thus facilitate dedicated fleet expansion without disruptions in service or compromises in service quality.

Varying levels of deployment throughout a contract period also has implications for planning and related functions, particularly in the formulation of RFPs and operating contracts. But such issues can be addressed by creating “sliding scales” which allocate the economies of scale to different fleet sizes, as vehicles and vehicle hours of service ebb and flow throughout the contract period. Such a mechanism not only effects equity in provider selection, but provides the lead agency with considerable latitude in effortlessly modifying the mix of dedicated and non-dedicated services throughout the contract period, with no significant inequities in compensation.

Sorting Trips

The levels of trips provided by each mode of service obviously reflects the numbers and types of trips sorted into each mode. This sorting is generally done by either a lead agency, a third party contractor or broker, or by the dedicated service provider. Permitting a contractor which provides both forms of service to make these decisions is obviously naïve, and will not likely yield the intended results. Regardless of which party performs the reservations and scheduling, and how it communicates with dispatchers, the general goal is to optimize performance by sorting as many trips as possible into dedicated service commensurate with safety, reliability and performance. As noted, measuring the performance results of this sorting process is virtually effortless.

An important corollary for accomplishing these broad goals is the need to control dedicated vehicle deployment hours and work shifts. Achieving this goal often requires the assignment of a large number of split-shifts, starting and ending at unusual times, all or many of which might not have existed in any form before the non-dedicated vehicles were introduced into the system (replacing underutilized segments of certain dedicated runs, from day to day). At the same time, the use of non-dedicated services can reduce the number of irregular and/or undesirable shifts – although a variety of shift times, lengths and formats is generally useful, as it provides drivers with more choices. Regardless, within this framework, the proper allocation of trips between dedicated versus non-dedicated services will mostly reflect the comparative costs of providing them by each mode. In most cases, it will generally make more sense to provide trips with non-dedicated vehicles than to deploy dedicated vehicles with gaps in their schedules to provide the same group of trips – depending somewhat, of course, on service costs and rates. The corollary of this principle is that dedicated vehicles which are less productive than non-dedicated vehicles should be removed from service during those times when they are.

Particularly if non-dedicated service is introduced as a new mode of service (rather than its roles simply expanded or modified), it is also not necessary – and may be problematic – to try to optimize performance of the mix too quickly. Instead, dedicated routes can be left “loose,” and trips more challenging to integrate into them can gradually be integrated as existing dedicated schedules are adapted to incorporate them, and drivers become increasingly familiar with the routes and their tightness. The reality is that schedulers can “cherry pick” those trips which can most easily be squeezed into dedicated service (again, commensurate with safety and reliability), and gradually move them around (within the 60-minute negotiation window allowable under the ADA) to permit the integration of more and more dedicated trips alongside them. Those trips excluded – the “odds and ends” – would then be provided by non-dedicated services. This refinement would not only occur over a period of time, but would embody another important principle of paratransit system design and operations: Every new trip scheduled provides an opportunity for performance improvement. Thus, the operating rationale in a mixed-mode system is to constantly refine performance, and constantly refine the mix of modes used to achieve it.

Non-Dedicated Target Trips

In a mixed-mode model, one must ask, “What will the non-dedicated component look like, and what percent of total trips will it likely involve?” As a starting point, in a large to moderately-sized urban/suburban service area with a moderately-sized taxi fleet, one can envision a system starting with about 65 to 70 percent dedicated service, and after a few months, ending up with roughly 80 percent. As noted, the crystal ball is unneeded, because the mix will “fall out” from the nature of the exercise outlined above. Service areas atypical of this model will likely require different mixes:

  • Large, mostly suburban or rural service areas with small fleet and user densities and long trip lengths may rely mostly on dedicated vehicles.
  • Dense urban areas with large fleet and user densities, and many short trips, would likely lean toward much greater reliance on taxis and other non-dedicated services, while dedicated vehicles would focus largely on the “cream” – group trips provided largely on a subscription basis to agency-affiliated clients, or relatively short and easily-prescheduled shopping, nutrition, and social service agency trips.
  • At the trip level, the basic goal is to pack as many safe, reliable trips as possible into dedicated vehicle runs, and “throw off” the overflow to non-dedicated providers – again, depending somewhat on relative costs and rates. Because their levels of service ebb and flow, because CPS trips simply thicken the densities of non-dedicated services, because their large fleets have numerous vehicles to draw upon for these assignments, and because last-minute assignments are the briar patch of non-dedicated services, their marriage with dedicated services is an ideal one for complementary paratransit service – even though the institutional and implementation issues and activities may produce a rocky honeymoon.

Complexity Versus Ease

Discussion of changes in service structure are common in communities providing only one mode of CPS. Union officials may recoil, board members may not understand the implications, lead agency decision-makers must rethink management and control, policy makers may be reluctant to modify regulatory and enforcement approaches (such as the willingness to replace non-compliant or recalcitrant taxi franchises), and drivers may be concerned that more methodical approaches may improve equity among drivers, and more likely change their hours of assignment – a paradigm shift with both losers and winners. In contrast, reservations, scheduling and dispatching personnel are likely to embrace the notion of introducing non-dedicated services like the monkeys at the conclusion of The Wizard of Oz. Suddenly unburdened by the labyrinth of arbitrary and counter-operational constraints, they would be free to not only employ their talents and creativity, but to do so in a rational operating environment with simple goals and the tools and services available to easily and logically meet them. FTA audits of many dedicated vehicle-only systems replete with multiple ADA violations have reflected such sensibilities.

Reliability Improvements and Implications

Regarding mythology that non-dedicated service is less reliable than dedicated service, such a belief reflects the quality of lead agency oversight and enforcement (as well as the willingness of elected officials to authorize them) far more than it does the quality of individual drivers or their management. Because monitoring of even dedicated service is generally poor or worse (this theme also emerged in virtually every securement-related lawsuit in which this author participated as an expert), even greater efforts will be needed to provide adequate monitoring of non-dedicated services. The most important element of these improvements is the regular review of drivers’ logs. In the case of non-dedicated services, their drivers must not only complete logs which include the full array of information needed by lead and operating agency personnel, but this information must be included, on the same logs, for non-CPS clients as well. Apart from the safety and liability issues these requirements involve, this review is a fundamental tool in the refinement of the mix of modes, and the only sure way to identify the genuine opportunities for performance improvement related to each individual trip.

Software Requirements and Considerations

The development and use of software for mixed-mode systems must change – and such changes may not be favored by software developers. Not only must algorithms be adjusted to perform new trade-offs, but the same level of reliance on these tools altogether would not be needed: A greater percentage of dedicated vehicle trips would be provided on a subscription basis, while more immediate-response trips would be provided by the non-dedicated fleet – with mode splits occurring constantly throughout the day to reflect both gaps in schedules (from late cancellations and “no-shows”) and counter-gaps (i.e., the evaporation of capacity from vehicles running behind schedule).

Within this environment, far more trips would be improvised – and most easily adjusted manually – often involving scheduling “overrides” at the digital level. And especially in extremely large systems, subscription rosters free from intrusion via the shoehorning of incompatible trips could be composed manually just as easily as digitally. The software would, instead, focus largely on trip-grouping queries, and on the rationalization of pre-scheduled demand responsive and immediate-response trips. Finally, most non-dedicated service could be handled digitally – if the contractor were capable. (Taxi dispatching involves the most rudimentary form of transportation algorithm, since it primarily identifies the vehicle nearest to the trip origin at a given point in time, with respect to its needed deadhead time.) At the same time, non-digital, non-dedicated service assigned to support or rescue dedicated service could be dispatched manually – since the complexities of the TX process would be simplified by fragmenting the clients into separate vehicles with no concurrent trip responsibilities.

In summary, the importance of digital technologies would change little; in truth, they would become more advanced, particularly in their application. But they would not perform the same roles as those which they presently do. The costs and challenges of software development to support such a model would be offset by the corresponding ease in performing key operating functions, and the cost savings which would accrue from their exercise. As a practical matter, some software packages already possess the capabilities to perform these new roles.

Additional Considerations

The problems to which an inappropriate mix of paratransit modes has contributed have also been the subject of more than one class action lawsuit (Richman v. SEPTA, Anderson v. Rochester-Genesee Regional Transportation Authority), while indirectly related to many others (Liberty Resources v. SEPTA, Flores v. LACMTA). Similarly, optimizing the mix of dedicated and non-dedicated services – much less merely including both modes – has been the subject of a number of FTA audits, as well as key debates in the redesign of a number of large North American paratransit systems (State of Rhode Island’s The Ride, Edmonton Transit DATS).

The most resistant efforts to maintain the status quo often yield either exponential costs or a significant level of trip denials. The latter generate class action suits, and the tenets of the systems’ ride-limiting barriers are slowly eroded. Most importantly, those systems still in violation risk losing dramatically more funds thorough the que tam process than they likely saved through years of suppressed costs through the employment of ride-limiting barriers like the provision of only dedicated vehicle service – and the multipliers of unmet trip demand that emerged in the process.

There is also a tendency in the provision of transportation, and particularly in the provision of demand-responsive service, to ignore or undervalue the importance of performance and reliability to safety and liability. In Today’s operating environment, the latter translate into significant operating costs. So it is noteworthy to point out that in this author’s examination of more than 50 wheelchair securement-related personal injury lawsuits, the vehicle was running behind schedule in almost every instance.

Curiously, the need for some non-dedicated serviced has even been recognized by the pupil transportation industry. At least two major systems – Detroit and Chicago – deploy taxicabs for both general and special education purposes – even though none of the vehicles deployed are even school bus-certified as per FMVSS requirements. (Conversely, 30 percent of the ridership on the Arcadia Dial-A-Ride (a general public system serving as the community’s local transit system since the mid-Seventies) are special education students taking home-to-school trips.) Regardless, the Detroit and Chicago experiences illustrate the potential for increased coordination or consolidation which the inclusion of non-dedicated services presents.

The mix of dedicated versus non-dedicated services must also reflect the often-peculiar, “unofficial” operating realities of the services involved. In communities with significant peak-hour traffic and yet no peak-hour variation in rates, many taxi drivers go off-duty during these hours, depleting the availability of service to the general public when it is most needed. Without serious enforcement efforts likely to be resented by drivers, the notion that many drivers would willingly shift their priorities to include non-tipping passengers who take longer to board and alight, who may occasionally not even know where they are going, and who’s directness-of-routing might be monitored by a public agency, is remote.

In worst-case scenarios, management must be more cautious, clever and vigilant, and far less naïve; taxi-based systems issuing discounted scrip are commonly defrauded by passengers, drivers and owners operating in concert: As an example, passengers may purchase $80 worth of scrip for $20. They then sell the scrip to drivers for $30. Driver’s turn them into their management for $40 (drivers commonly receive “half the take”). Finally, management submits the scrip to the lead agency for the full $80. In the final analysis, it cost the lead agency $60 to enrich three layers of crooks, for whom not a single trip was provided.

Further Study on the Horizon

At the time of this document’s creation, the Transportation Research Board’s Transportation Cooperative Research Program (TCRP) had already issued a problem statement on this subject, announcing the imminent issuance of a Request for Proposals to examine the subject in greater detail, including the development of algorithms to facilitate a determination of the optimum mix of service modes – presumably at both the planning and operating levels. While one may perceive this study as placing several carts before the proverbial horse, it must also be acknowledged that decision-making in most ADA-era paratransit systems is articulated, if not governed, by digital applications. If scheduling of small, dedicated-vehicle-only systems is performed by using software, it is unrealistic to expect management to revert to manual scheduling even when modifications like the inclusion of non-dedicated services makes it far easier to do so.

Largely for this reason, the TCRP project is of considerable importance to the transit community, and should yield an invaluable set of information and tools for improving system performance consistent and compatible with the manner in which most paratransit staff currently reserve, schedule and dispatch trips.

Conclusions

As noted, the availability of dual modes with complementary capabilities will greatly simplify the provision of service as well as improve its results. In the simplest example, every fifth dedicated trip might involve difficult integration, may create hardships for four otherwise-pleased passengers, increase their ride times, undermine performance and, in simple terms, bastardize an otherwise logical, quick, reliable and efficient trip segment. Simply tossing that one trip onto the non-dedicated over-flow “system” (which also provides better service for the ungrouped client) transforms a puzzle with thorns into a simple, easy, efficient and reliable journey for all involved, while reducing the costs of providing these trips borne largely by those not taking any of them (i.e., non-ADA-eligible taxpayers).

Meeting ADA requirements can certainly be accomplished without employing non-dedicated vehicles. But doing so will irrefutably cost more, complicate the system’s operations, require a flotilla of costly dedicated vehicles- and drivers-in-waiting to cover fluctuations in demand, waste significant opportunities for performance improvement, and keep the trip reservation, scheduling and dispatching processes from becoming rational, easy and creative instead of stressful, chaotic and dysfunctional, and fraught with mental gymnastics, frustration, dissatisfaction and resentment.

Regarding single-mode approaches to the provision of paratransit service, the costs for providing the last increment of service with a dedicated vehicle is exponentially greater than those associated with the provision of the “average” trip, much less the average trip which remains after the “bad ones” have been removed from the mix. The further these costs escalate, the further isolated they become from a rational and responsible solution to operations, and the greater they become a disruption to it.

Finally, the full maximization of performance by optimizing the mix of dedicated and non-dedicated service is likely to occur only with a complete reversal in transit and other lead agency philosophy: The practice of controlling costs by limiting access to service must be replaced with the practice of providing service as efficiently as possible. In truth, a decade-plus of deterring ridership through the employment of ride-limiting barriers probably reduced costs more than would have the optimization of performance. But as more and more eligible clients are introduced to complementary paratransit service – especially following changes made in response to FTA audits, the involvement of local advocacy groups, and the filing of class action suits – this demand can reasonably be expected to grow, and grow substantially. At some point, the switch to a philosophy of providing these trips as efficiently as possible, in accordance with the principles outlined above, will result in superior performance and lower costs than have traditional approaches.

The irony of this issue – particularly the need to “define” the optimum mix of both forms of service – is that the essence of these modes’ characteristics necessarily establishes this optimum as a simple bi-product of their operating dynamics: The very nature of non-dedicated services is that they ebb and flow in response to demand and usage. Over time, the optimum mix will become obvious, and will seek its own level, even without much thought or methodology. This revelation would emerge from the simple measurement of trial-and-error efforts to assign trips to dedicated versus non-dedicated vehicles, and the selection of a crude starting point. But the tools envisioned by the TCRP for facilitating this refinement will make the process easier, quicker, less costly, and more accessible to a broad range of paratransit officials, management and staff.