Structural Determinants of Pest Management Failure in Attached Residential Housing

1. Introduction: failure as a structural property

Ask a pest management professional why a bed bug programme failed in an apartment building and the answers cluster into a familiar set: the resident did not prepare the unit, the neighbouring tenant refused access, the landlord would not authorise adjacent inspections, the budget covered one suite. Each of these is true in individual cases. Collectively they describe a pattern so consistent that it should be understood not as a series of unfortunate operational lapses but as the predictable output of a structurally mismatched intervention.

The mismatch is one of scale. A dwelling unit is a legal and financial object. It is the thing a lease attaches to, the thing a work order is written against, and the thing an invoice is addressed to. It is not, however, a biological boundary. The evidence assembled over the past fifteen years demonstrates that for Cimex lectularius in particular, and, with variation, for Blattella germanica, Monomorium pharaonis and commensal rodents, the functional population unit in an attached building is the building.

This paper sets out the evidence for that claim, traces its mechanisms through building physics and pest behaviour, examines the Manitoba legal framework that in principle already requires a broader scope, and quantifies why the economics favour building-wide intervention even for a landlord acting purely in self-interest. It closes with a specification of what a defensible programme actually requires, and an honest account of where the evidence runs out.

1.1 Scope and definitions

Throughout, attached housing denotes any residential building in which dwelling units share at least one structural element, a party wall, a floor-ceiling assembly, a common corridor, a shared mechanical chase, or a contiguous roof void. This includes low-rise walk-ups, high-rise apartment blocks, townhouse and rowhouse rows, converted single-family houses, and the three-storey wood-frame stock that constitutes a substantial share of Winnipeg's older rental inventory.

We distinguish active dispersal (the pest moves under its own power between units) from passive or human-mediated dispersal (the pest is carried between units on belongings, furniture, laundry, or persons). Both operate in attached housing; they have different mitigations and are frequently conflated in practice, to the detriment of both.

2. The dispersal evidence base

The single most consequential body of evidence for multi-family pest management is the set of field studies conducted in occupied apartment buildings from roughly 2010 onward. Prior to this work, inter-unit spread was widely assumed to be predominantly passive, a matter of residents carrying infestations between units. The field data substantially revised that assumption.

2.1 Interceptor surveillance and the scale of spread

Wang and colleagues conducted longitudinal surveillance in a 223-unit high-rise complex, using interceptor traps placed both inside apartment entry doors and in the common hallway immediately outside them.¹ The findings are worth stating in full because they establish the quantitative baseline for everything that follows:

Two of these figures deserve emphasis. The first is the 53% adjacency rate: if a unit borders a known infestation, the prior probability that it is itself infested is approximately a coin flip. This is not a marginal elevation in risk; it is the single strongest predictor available to a programme designer, and it is available before any inspection is conducted.

The second is the 50% unawareness rate. This figure alone falsifies the operating assumption of complaint-driven pest management. If half of infested residents do not report, whether because they are asymptomatic to bites, attribute the bites to another cause, fear eviction or stigma, or simply have not yet noticed, then a programme that responds only to complaints is, by construction, addressing at most half of the population it is nominally managing.

2.2 Mark-release-recapture and movement distances

Cooper, Wang and Singh used mark-release-recapture across six infested apartments to characterise movement directly rather than inferring it from aggregation patterns.² The method matters: earlier work had largely relied on the spatial distribution of harbourages, which cannot distinguish a bug that travelled from one that was always there. Marking individuals and intercepting their movement over 32 days resolved that ambiguity.

The population estimates the study generated are, on their own, an argument for a different programme design. Estimated bed bug populations in the monitored apartments ranged from 2,433 to 14,291 individuals at four to seven days post-release.² These are not the numbers implied by a service model that budgets ninety minutes and two follow-up visits per unit.

The study's operational recommendation was explicit and is still not universally implemented: neighbouring-unit inspection should be expanded beyond the conventional above-below-adjacent triad to include units across the hallway from a known infestation.² The mechanism is the entry-door corridor pathway documented in the interceptor work, bugs move to the door area, exit into the hallway, and enter whatever door they next encounter, which is as likely to be opposite as beside.

2.3 Population genetics as independent corroboration

Behavioural and trapping studies establish that movement occurs. Population genetics establishes what that movement has actually produced over time. Saenz and colleagues developed high-resolution microsatellite markers and applied them to infestation dynamics across three multi-storey apartment buildings.³

The genetic signal was consistent: populations within single apartments showed high relatedness and low diversity, the signature of founding from a small propagule, while the relationships between apartments indicated extensive within-building spread. In one building the entire infestation traced to a single introduction followed by dispersal; in others, two or more introductions were evident.³

This is an important independent line of evidence because it is not subject to the observational limitations of trapping. A trap tells you a bug moved during your study window. Genetics tells you what the cumulative history of movement has been, and in these buildings that history was one of building-scale connectivity.

2.4 Longitudinal spread in institutional housing

A complementary data point comes from medical school housing, where an infestation spread from a single unit to 68 units over a 25-month period.² In both this case and the 223-unit study, more than half of infested units shared a common wall, floor or ceiling with another infested unit.¹²

The institutional context is instructive because it controls somewhat for the socioeconomic variables that confound apartment studies. This was not a building where residents lacked resources to respond, nor one where reporting carried the eviction risk that suppresses disclosure in private rental markets. The spread happened anyway, which suggests the mechanism is structural rather than behavioural.

3. Structural pathways in the built environment

The dispersal literature establishes that movement occurs and quantifies its rate. It says less about the specific architectural features that mediate it. This section addresses that gap, with particular attention to construction types prevalent in Winnipeg's rental stock.

3.1 The party wall assembly

In wood-frame multi-family construction, the party wall between units is not a solid barrier. It is an assembly, typically gypsum board over studs, with a cavity that may be insulated, partially insulated, or open. That cavity is continuous over substantial areas and is penetrated at every electrical box, plumbing stub, low-voltage run and mechanical chase.

Electrical outlet boxes deserve particular attention. In older construction, back-to-back outlets sharing a stud cavity create a direct aperture between units with no intervening material at all beyond the box knockouts. A harbourage behind a baseboard in one unit and a harbourage behind the adjacent baseboard in the neighbouring unit may be separated by a few centimetres of air.

The patent literature is unusually candid on this point, noting that inter-unit spread depends in part on the material used to partition units, with concrete constituting a more effective barrier than lighter assemblies.⁴ This is a useful practical heuristic: a concrete-plank high-rise and a wood-frame walk-up of the same unit count are not equivalent risk environments, and a programme designed for one may be miscalibrated for the other.

3.2 Floor-ceiling assemblies and vertical pathways

Vertical spread is mediated by plumbing stacks, mechanical chases, and the floor-ceiling cavity itself. In balloon-framed older housing, including many Winnipeg conversions, stud cavities may run continuously from basement to roof void, providing an uninterrupted vertical pathway that bypasses every horizontal barrier in the building.

The practical consequence is that the conventional inspection triad of above, below, adjacent is a reasonable first approximation but is not derived from the building's actual connectivity graph. A unit two floors up on the same plumbing stack may be more closely connected, in dispersal terms, than the unit immediately next door on the other side of a well-sealed fire-rated assembly.

3.3 Common corridors as transmission surface

The entry-door pathway documented by Wang et al. reframes the common corridor from circulation space to epidemiological surface.¹ Bed bugs move to the door area, disperse into the hallway, and enter other units. Carpeted corridors with substantial baseboard gaps provide both harbourage and a travel route; the corridor is not merely a conduit but can itself sustain a population.

This has an operational implication that is frequently missed: corridor monitoring is a surveillance opportunity, not merely a treatment target. Interceptors in corridors detect movement that unit-interior monitoring misses, and they do so without requiring resident consent or unit access, which, as discussed in §7, is often the binding constraint.

3.4 Shared mechanical and laundry systems

Central laundry facilities concentrate textile movement from every unit in the building through a single room. For Cimex this is both a risk and a control point. Common waste chutes, similarly, aggregate material from every floor and constitute a documented harbourage in cockroach and rodent management.

For Monomorium pharaonis, the pharaoh ant, shared mechanical systems are more than a pathway; they are habitat. The species requires sustained warmth in the region of 27–30 °C, a condition reliably supplied by heated mechanical rooms and service chases in Canadian buildings year-round, decoupling the population entirely from outdoor seasonality.

4. The unit-scope error

We can now state the failure mechanism precisely. Consider a building in which unit i is reported infested and treated to a high professional standard, achieving complete elimination within that unit. Let p be the probability that at least one adjacent unit harbours an untreated population, and let d be the inter-unit dispersal rate.

From the surveillance data, p ≈ 0.53 for a single adjacency,¹ and rises rapidly with the number of shared boundaries. A mid-floor, mid-corridor unit may share boundaries with four to six other units once across-hallway connectivity is included.² Under independence, a conservative assumption, since adjacency infestations are positively correlated through shared pathways, the probability that no neighbouring unit is infested falls below 0.05 at five adjacencies.

The recolonisation timescale follows from the measured dispersal rate of approximately six individuals per apartment per four weeks moving through entry doors.¹ A gravid female is sufficient to found a new population. The mean time to reintroduction is therefore on the order of weeks, comfortably shorter than the two-to-four-week follow-up interval of a conventional programme, and far shorter than the interval before the next complaint is filed.

4.1 Why this presents as treatment failure

From the resident's perspective, the sequence is: treatment, brief relief, return of bites. From the landlord's perspective: an invoice, a quiet period, a repeat complaint. Both parties conclude that the treatment failed. The contractor, knowing the application was competent, concludes that the resident reintroduced the infestation or failed to prepare properly.

All three interpretations are wrong in the same way. The treatment succeeded at its stated scope and the scope was wrong. This misattribution is corrosive: it damages the contractor relationship, encourages residents to attempt DIY intervention with the consequences described in §6, and frequently produces a switch to a cheaper contractor, which changes nothing, because the binding constraint was never contractor quality.

4.2 The comparative evidence

The clearest experimental statement of this problem comes from Wang and colleagues' comparison of three management strategies in a low-income apartment building.⁵ The study evaluated non-chemical methods, insecticide-only treatment, and integrated pest management combining both. Its conclusion regarding residual infestations is directly on point: the presence of low numbers of bed bugs in some test apartments at study end was attributable to dispersal from neighbouring infested units and/or passive dispersal, and, critically, the slow elimination process in test apartments might itself have produced new infestations in adjoining units.⁵

The authors' recommendation is unambiguous: building-wide intervention is important for bed bug management in multi-unit dwellings.⁵ They further note that education reduces human-mediated passive dispersal but cannot prevent active dispersal, because numerous pathways exist between apartments regardless of resident behaviour.⁵ This last point is worth dwelling on, because resident education is the most common response to repeated failure and it addresses only one of the two dispersal modes.

5. Detection lag and the asymptomatic reservoir

If half of infested residents are unaware of their infestation,¹ then at any given moment a complaint-driven programme has an unobserved reservoir approximately equal in size to its observed caseload. This section examines why.

5.1 Sources of detection lag

Dermatological variation. Reaction to Cimex bites varies widely between individuals, from prominent wheals to no visible response at all. A resident who does not react is not an unusual case; they are a substantial minority of the population and a reliable source of undetected infestation.

Misattribution. Bites are commonly attributed to mosquitoes, fleas, spiders, or dermatological conditions. In Winnipeg, where Aedes vexans pressure is a fact of summer life, summer-onset bed bug bites have a highly available alternative explanation.

Population size at onset. A founding population of one gravid female produces detectable bite frequency only after several generations. At the roughly five-to-seven week egg-to-adult interval typical of heated buildings, the interval between introduction and reliable detectability is measured in months.

Disclosure suppression. In private rental markets, reporting carries perceived risk: blame, charge-back, lease non-renewal, or social stigma. This is not irrational; some of these outcomes do occur. The result is a reporting rate below the true detection rate.

5.2 The implication for surveillance design

A programme that relies on complaints as its detection mechanism has a sensitivity ceiling set by these factors, and that ceiling appears to be near 50%.¹ No improvement in treatment quality can compensate for a surveillance system that misses half of cases, this is a detection problem, not a treatment problem, and it requires a detection solution.

Passive interceptor monitoring is the obvious candidate. Interceptors detect presence independent of resident reporting, independent of dermatological reaction, and at population levels well below the threshold of subjective awareness. The literature uses them as research instruments precisely because they are more sensitive than the alternatives.¹⁵

6. Chemical amplification of dispersal

The preceding sections describe dispersal as a background process. Certain interventions accelerate it substantially, and the most common of these is the application of repellent insecticide by residents or by inadequately specified contractors.

6.1 The repellency mechanism

Pyrethroid formulations sold in the consumer channel are typically repellent at field concentrations. Applied to a harbourage, they do not reliably kill the population; they render the harbourage aversive. The behavioural response is dispersal away from the treated surface, into wall voids, along the pathways described in §3, and into adjacent units.

This is not a subtle effect. It converts a spatially contained infestation into a distributed one, and it does so on a timescale of hours. A single resident's aerosol application can undo the containment on which a building-wide programme depends.

6.2 The pharaoh ant case

The most extreme instance of chemically induced spread is Monomorium pharaonis. Colonies of this species respond to chemical stress by budding: queens and workers separate and establish satellite colonies elsewhere in the structure. Repellent application against pharaoh ants does not reduce colony count; it increases it. In a multi-family building this converts a single-unit problem into a distributed, building-wide one that is materially harder and more expensive to resolve than the original.

This is why healthcare facility protocols specify baiting-only approaches for pharaoh ants, and why the same logic should apply in multi-family residential settings, where the structural connectivity is comparable and the supervision is weaker.

6.3 Resistance as a compounding factor

Widespread consumer pyrethroid use in multi-family housing also constitutes an uncontrolled selection regime. Sub-lethal exposure across a connected metapopulation is close to a textbook recipe for resistance development, and the resulting resistant population is shared by every household in the building. The externality is unpriced and, under current arrangements, unaddressed.

7. The Manitoba statutory framework

Manitoba's legal position on this question is, in outline, already aligned with the evidence, a fact not widely appreciated by either landlords or tenants.

7.1 Allocation of responsibility

Under Manitoba's residential tenancies framework, pest control is a landlord responsibility. The Residential Tenancies Branch identifies mice, cockroaches, bed bugs and other pests explicitly as health risks that landlords are obliged to keep under control.⁶ The obligation attaches to the landlord because the landlord controls the building, and, as this paper argues, the building is the relevant unit of intervention.

7.2 Access and notice

The practical objection to building-wide intervention is access: neighbouring residents may decline inspection. Manitoba's framework anticipates this. Landlords may enter units for inspection and treatment on written notice, subject to a minimum of 24 hours and a maximum of two weeks.⁶ Access is therefore a procedural requirement rather than a veto, provided notice is given correctly.

7.3 The cluster inspection requirement

Most significantly, Manitoba guidance specifies that when a unit is treated for bed bugs, the units beside, above and below must also be inspected.⁶ Single-suite response is, in the provincial framing, inadequate.

Two observations follow. First, the practice gap: single-unit treatment remains common despite this guidance, which suggests the guidance is either unknown or unenforced. Second, the evidence gap in the other direction: the specified triad omits across-hallway units, which the mark-release-recapture data identify as a genuine pathway.² Manitoba's requirement is a floor, not a ceiling, and a programme designed to the evidence should exceed it.

8. Economics of the failed treatment

The standard objection to building-wide intervention is cost. This section argues that the objection inverts the actual economics.

8.1 The cost of repetition

Let C_u be the cost of treating a single unit and n the number of repeat treatments before the problem is either resolved or abandoned. Under the recolonisation dynamics of §4, n is not small: each cycle achieves temporary local elimination followed by reintroduction. Total expenditure is nC_u with a terminal state of continued infestation.

Building-wide intervention across m units costs approximately mC_u discounted by the substantial per-unit efficiency of treating contiguous units in one mobilisation , no repeat travel, no repeat set-up, shared inspection time. The comparison is therefore between nC_u with failure and roughly 0.6mC_u with resolution. For a building where the infestation has reached the 45% prevalence documented in the surveillance literature,¹ the repeat-treatment path is straightforwardly more expensive.

8.2 Unpriced costs

The direct treatment invoice understates total cost considerably. Additional burdens include tenant turnover and vacancy loss; furniture disposal and replacement, frequently borne by residents least able to absorb it; Residential Tenancies Branch proceedings and associated administrative time; reputational effects in a rental market where infestation history circulates informally; and, for residents, the documented psychological burden of chronic infestation.

None of these appear on a pest control invoice, and all of them scale with the duration of the infestation rather than with the number of treatments. An intervention that resolves in one mobilisation rather than persisting for years dominates on total cost even where its line-item cost is higher.

8.3 Incentive misalignment

A structural note: the contractor paid per unit treated has no financial interest in the scope correction advocated here, and the landlord authorising treatment per complaint has no visibility into the reservoir. The arrangement produces exactly the behaviour observed. Correcting it requires either a building-wide service contract, which aligns contractor and landlord incentives toward elimination, or a regulatory requirement with genuine enforcement.

9. Specification for a defensible programme

The preceding analysis supports a concrete specification. A multi-family programme is defensible if it satisfies the following.

9.1 Surveillance

• Passive monitoring independent of complaints. Interceptor devices in a defined sample of units, with coverage weighted toward known-infested clusters, addressing the 50% unawareness rate directly.¹
• Corridor monitoring at entry doors. Detects inter-unit movement and requires no unit access.¹
• Baseline building survey at programme initiation, rather than inferring extent from complaint history.

9.2 Scope

• Cluster inspection exceeding the statutory triad: beside, above, below, and across the corridor.²⁶
• Connectivity-informed extension where the building's plumbing stacks or balloon framing create non-obvious pathways (§3.2).
• Simultaneous rather than sequential treatment of a cluster, so that treated units are not recolonised from untreated neighbours mid-programme.⁵

9.3 Chemistry

• Non-repellent actives and baits as the default, to avoid the dispersal amplification of §6.
• Baiting-only protocol for pharaoh ants, without exception.
• Resident communication that specifically warns against consumer aerosols, framed as protection of neighbours rather than as instruction.

9.4 Documentation

• Dated records adequate for Residential Tenancies Branch proceedings, protecting the landlord's due-diligence position.⁶
• Building-level trend data rather than per-unit service tickets, so that prevalence can be tracked and the programme evaluated on elimination rather than on activity.

10. Limitations and open questions

Intellectual honesty requires stating clearly what this analysis does not establish.

Geographic transferability. The core field studies were conducted in New Jersey, North Carolina and comparable jurisdictions.¹²³ Winnipeg's building stock, tenancy law and climate differ. The dispersal mechanisms are architectural and behavioural rather than climatic, which supports transferability, but no equivalent Manitoba surveillance study has been published. This is a genuine gap and we would welcome its closure.

Species generalisation. The quantitative dispersal data concern Cimex lectularius. The structural argument extends naturally to Blattella germanica and Monomorium pharaonis on mechanistic grounds, but the specific rates differ and should not be transferred numerically.

Construction-type stratification. The observation that concrete partitions resist spread better than lighter assemblies⁴ is directionally sound but poorly quantified. A stratified analysis of spread rate by construction type would materially improve programme design and does not appear to exist.

Cost modelling. The economic argument in §8 is structural rather than empirical. We are not aware of a published cost comparison of unit-scoped versus building-wide intervention with adequate follow-up, and the efficiency discount applied is an operational estimate rather than a measured figure.

11. Conclusion

Pest management failure in attached housing is not primarily a failure of products, applicators or residents. It is a failure of scope, and it is predictable from first principles once the dispersal evidence is taken seriously.

The evidence is not ambiguous. A single introduction reached 45% of a 223-unit building in 41 months.¹ Units adjacent to an infestation are themselves infested 53% of the time.¹ Half of infested residents do not know.¹ Population genetics confirms extensive within-building spread from small founding propagules.³ The authors of the principal comparative trial state plainly that building-wide intervention is necessary.⁵

Manitoba's statutory framework already requires more than single-unit response, specifying inspection of units beside, above and below a treated unit.⁶ The gap is one of practice, and partly one of ambition: the evidence supports a scope somewhat wider than the statute specifies.

The practical recommendation is narrow and, we would argue, uncontroversial: stop treating the dwelling unit as the epidemiological boundary of an infestation that is demonstrably building-wide. Every downstream problem in multi-family pest management, the repeat callbacks, the blamed residents, the exhausted budgets, the tenancy disputes, follows from that single category error.

References

1. Wang, C., Saltzmann, K., Chin, E., Bennett, G.W. & Gibb, T. (2010). Characteristics of Cimex lectularius (Hemiptera: Cimicidae) infestation and dispersal in a high-rise apartment building. Journal of Economic Entomology, 103(1), 172–177. Summary of findings and interceptor methodology discussed in Wang, C. & Cooper, R., "Understanding Bed Bug Infestation and Dispersal Patterns," Pest Control Technology. https://www.pctonline.com/article/pct1003_understandingbb/
2. Cooper, R., Wang, C. & Singh, N. (2015). Mark-Release-Recapture Reveals Extensive Movement of Bed Bugs (Cimex lectularius L.) within and between Apartments. PLOS ONE, 10(9), e0136462. https://journals.plos.org/plosone/article?id=10.1371%2Fjournal.pone.0136462
3. Saenz, V.L., Booth, W., Schal, C. & Vargo, E.L. (2012). Genetic analysis of bed bug populations reveals small propagule size within individual infestations but high genetic diversity across infestations from the eastern United States. Journal of Medical Entomology, 49(4), 865–875. Microsatellite development and within-building infestation dynamics. https://u.osu.edu/bedbugs/research-refs/bed-bug-dispersal/
4. Methods, apparatus and compositions for abatement of bedbugs. United States Patent 9,686,973. Discussion of partition material and its effect on inter-unit spread. https://image-ppubs.uspto.gov/dirsearch-public/print/downloadPdf/9686973
5. Wang, C., Gibb, T. & Bennett, G.W. (2012). Comparison of Three Bed Bug Management Strategies in a Low-Income Apartment Building. Insects, 3(2), 402–409. doi:10.3390/insects3020402 https://doi.org/10.3390/insects3020402
6. Manitoba Residential Tenancies Branch. Landlord and tenant responsibilities regarding pest control, entry and notice requirements, and inspection of units adjacent to a treated unit. Province of Manitoba. https://www.gov.mb.ca/cca/rtb/