UBC Faculty Research and Publications

Protocol of the impact of alternative social assistance disbursement on drug-related harm (TASA) study… Richardson, Lindsey; Laing, Allison; Milloy, M-J; Maynard, Russ; Nosyk, Bohdan; Marshall, Brandon David Lewis; Grafstein, Eric; Daly, Patricia; Wood, Evan; Montaner, Julio; Kerr, Thomas 2016

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-Richardson_L_et_al_Protocol_impact_alternative_social.pdf [ 717.99kB ]
JSON: 52383-1.0343042.json
JSON-LD: 52383-1.0343042-ld.json
RDF/XML (Pretty): 52383-1.0343042-rdf.xml
RDF/JSON: 52383-1.0343042-rdf.json
Turtle: 52383-1.0343042-turtle.txt
N-Triples: 52383-1.0343042-rdf-ntriples.txt
Original Record: 52383-1.0343042-source.json
Full Text

Full Text

STUDY PROTOCOL Open AccessProtocol of the impact of alternative socialassistance disbursement on drug-relatedharm (TASA) study: a randomizedcontrolled trial to evaluate changes topayment timing and frequency amongpeople who use illicit drugsLindsey Richardson1,2* , Allison Laing1, M-J Milloy1,3, Russ Maynard1,4, Bohdan Nosyk1,5, Brandon Marshall6,Eric Grafstein7,8, Patricia Daly8,9, Evan Wood1,3, Julio Montaner1,3 and Thomas Kerr1,3AbstractBackground: Government social assistance payments seek to alleviate poverty and address survival needs, but theirmonthly disbursement may cue increases in illicit drug use. This cue may be magnified when assistance isdisbursed simultaneously across the population. Synchronized payments have been linked to escalations in druguse and unintended but severe drug-related harms, including overdose, as well as spikes in demand for health,social, financial and police services.Methods/design: The TASA study examines whether changing payment timing and frequency can mitigatedrug-related harm associated with synchronized social assistance disbursement. The study is a parallel armmulti-group randomized controlled trial in which 273 participants are randomly allocated for six assistancecycles to a control or one of two intervention arms on a 1:1:1 basis. Intervention arm participants receivetheir payments: (1) monthly; or (2) semi-monthly, in each case on days that are not during the week whencheques are normally issued. The study partners with a community-based credit union that has developeda system to vary social assistance payment timing. The primary outcome is a 40 % increase in drug useduring the 3 days beginning with cheque issue day compared to other days of the month. Bi-weeklyfollow-up interviews collect participant information on this and secondary outcomes of interest, includingdrug-related harm (e.g. non-fatal overdose), exposure to violence and health service utilization. Self-reporteddata will be supplemented with participant information from health, financial, police and governmentadministrative databases. A longitudinal, nested, qualitative parallel process evaluation explores participantexperiences, and a cost-effectiveness evaluation of different disbursement scenarios will be undertaken.Outcomes will be compared between control and intervention arms to identify the impacts of alternativedisbursement schedules on drug-related harm resulting from synchronized income assistance.(Continued on next page)* Correspondence: uhri-lr@cfenet.ubc.ca1British Columbia Centre for Excellence in HIV/AIDS, St. Paul’s Hospital,608-1081 Burrard Street, Vancouver V6Z 1Y6, BC, Canada2Department of Sociology, University of British Columbia, 6303 NW MarineDrive, Vancouver V6T 1Z1, BC, CanadaFull list of author information is available at the end of the article© 2016 The Author(s). Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Richardson et al. BMC Public Health  (2016) 16:668 DOI 10.1186/s12889-016-3304-6(Continued from previous page)Discussion: This structural RCT benefits from strong community partnerships, highly detailed outcomemeasurement, robust methods of randomization and data triangulation with third party administrativedatabases. The study will provide evidence regarding the potential importance of social assistance programdesign as a lever to support population health outcomes and service provision for populations with a highprevalence of substance use.Trial registration: NCT02457949 Registered 13 May 2015.Keywords: Social assistance, Drug use, Drug-related harm, Structural interventionBackgroundSocial assistance, including disability support, incomeassistance or other state-provided cash transferbenefits, is commonly distributed on a monthly basis[1, 2], and can provide important protection againstthe harms associated with extreme poverty [2].Recipients of social assistance commonly increasegeneral consumption following benefit receipt [3].For people who use illicit drugs (PWUD), receipt ofsuch benefits may also serve as a cue for intensifieddrug use [4] that is magnified when receipt issynchronized across the population [5–7]. Previousobservational research has linked monthly social as-sistance payments to unintentional, yet cyclical andsevere increases in drug-related harms [7–9], as wellas increased demand for health, social, financial andpolice services [6–8, 10–13]. While there have beenrepeated calls for interventions to mitigate drug-related harms associated with synchronized social as-sistance [7, 8, 12], research to date has been limitedto observational studies and a single natural experi-ment. We have therefore designed and undertaken acontrolled, experimental study that examines whetherchanging the timing and frequency of social assistancecash transfers can reduce drug-related harm linked tothe synchronization of such payments.Socioeconomically marginalized PWUD commonly relyon social assistance as a critical source of income [14, 15].However, the temporal synchronization of such paymentshas significant consequences for PWUD, their socialcontacts and service providers. Specifically, previous re-search has linked payment timing to high intensity druguse [6, 10, 16], higher risk drug use [5], increased soberingor detoxification unit admissions [7, 17], drug-relatedemergency department use and hospitalization [11, 13, 16,18, 19], fatal and non-fatal overdose [5, 7, 12, 20, 21]hospital discharges against medical advice [9, 22], publicdisorder [7, 11], addiction and HIV treatment interruption[9, 22, 23], mental health apprehensions [24] and barriersto health service access [6]. Additionally, such increasesimpact demands on health, social, financial and policeservice provision. For example, providers have noted sig-nificant increases in patient volumes for harm reductionservices such as supervised injecting sites [6], emergencydepartment and psychiatric emergency services, policeservice calls and increases in cash outlays from financialservice providers located in inner-city neighbourhoods (M.Corral, J. Chu, J. Fahey, personal correspondence). Chal-lenges related to service provision are taxing for providers,interfere with their ability to provide quality and timelyservices and may result in individuals leaving service sitesbefore accessing critical supports [6].Initial documentation of drug use coinciding withcheque issue pointed to a “cheque effect”, where druguse was attributed to the provision of disabilityassistance [10]. A recent review examined whether dis-ability payments are the cause of increased illicit druguse overall or whether they simply alter the timing ofdrug use [8]. Findings found no difference in overallrates of drug use between those receiving or not receiv-ing disability benefits, but that payments and spikes indrug use intensity were linked. Delayed payments haveproduced corresponding delays in rates of drug-relatedhospitalization [8]. Studies provide a strong signal, first,that payments do not alter whether people use drugs,and second, that payment timing may be an importantlever through which to influence payment-related in-creases in drug use and consequent harm. In addition, astrong rationale for leveraging pay frequency in order tosmooth consumption patterns exists [25], although thishas not been explored among PWUD. Smaller, moreregular payments may further decrease or potentiallydisperse drug use and related harm. Despite indicationsthat changing the timing and frequency of paymentscould significantly benefit PWUD, their communitiesand service providers, there is no evidence from con-trolled experiments identifying the impacts of such astrategy. This is a critical gap given the health, social andeconomic costs of drug-related harm associated withcheque issue and potentially important implications forsocial assistance policy.Drug-related concerns coinciding with cheque issuemay be particularly acute in inner-city neighbourhoodswith higher prevalence of illicit drug use andconcentrations of social assistance recipients [26]. TheDowntown Eastside (DTES) neighbourhood in Vancouver,Richardson et al. BMC Public Health  (2016) 16:668 Page 2 of 16Canada is one such area, commonly characterized by: highrates of poverty; an open drug market; high prevalence ofillicit drug use, mental health disorders, and HIVinfection; and a large proportion of residents receivinggovernment assistance [27]. Cheque issue days, generallyon the last Wednesday of each month, are referred tolocally as “Cheque Day” or “Welfare Wednesday”. Thewidespread and severe harms produced by synchronizedsocial assistance are widely acknowledged among commu-nity members and service providers and have been docu-mented by over 20 years of ongoing observational research[7, 24, 28]. The identification of public health- andcommunity safety-promoting approaches to the dis-bursement of social assistance is therefore an urgentpublic health and community safety priority inVancouver as elsewhere, with significant implicationsfor policy and service provision.We have therefore designed and initiated a prospectiverandomized controlled trial (RCT), the impact of Alterna-tive Social Assistance disbursement on drug-related harm(TASA) study, to examine whether changing the timingand frequency of social assistance cash transfers canreduce drug-related harm linked to the synchronization ofsuch payments. The TASA study is a demonstration pro-ject of the British Columbia node of the CanadianResearch Initiative in Substance Misuse (CRISM). Thestudy changes the temporal conditions of social assistancereceipt to determine which disbursement arrangementmost effectively mitigates escalations in drug-relatedhealth, social and economic harm associated with socialassistance cheque issue days. Contrasted with behaviouralinterventions that seek to directly alter individual behav-iour, the study is a structural intervention that alters thecontext (i.e., socio-economic environment) in whichhealth risks are produced [29]. TASA focuses on both thetiming and frequency of income assistance payments bycomparing two distinct interventions. The first involvesthe disbursement of income assistance payments once amonth, as is currently the case. However, payments aremade on a day that does not fall during cheque issueweek, and participants are randomized to different days,rather than all intervention participants receiving paymenton the same non-government cheque issue day. Thisallows us to determine whether staggering paymentsseparates individual cues for drug use following individualpayment from the social cues for increased drug use thataccompany synchronized payments from all participantsbeing paid on the same day. The second interventioninvolves twice-monthly payments, similarly on days thatdo not fall during cheque issue week and not on the sameday for all participants. In addition to differentiatingbetween individual and social cues for drug use, this strat-egy tests whether splitting income assistance paymentscan reduce drug use by smoothing consumption patterns.Both interventions are administered through Pigeon ParkSavings, a branch of Vancouver City Savings Credit Unionin the DTES specifically designed to support residentswho face barriers to accessing services at other financialinstitutions.There are potentially important differences betweenthe two interventions that may affect their relative effi-cacy. For example, staggering social assistance paymentscould displace consumption for individuals away fromgovernment cheque issue, but may not reduce overallconsumption. At scale, this may improve experiences ofservice providers who would not have to manage cyclicalescalations in service demand, but may provide littlebenefit to individual PWUD other than improved serviceaccess. Staggering and splitting income assistance pay-ments may support smoothing consumption patterns, butmay potentially result in multiple periods of escalateddrug use signaled by more than one monthly payment.The testing of two different payment schedules enables usto examine which approach, if either, more effectivelydecreases spikes in drug use and drug-related harm coin-ciding with synchronized social assistance disbursement.ObjectivesThe main objective of TASA Cheque issue study is to as-sess the effectiveness of structural changes in the timingand frequency of social assistance payments in reducingdrug use and drug-related harm in the days surroundingsocial assistance cheque issue in a population of socio-economically marginalized people who use illicit drugs.The interventions will be compared with each other andwith a no-intervention control group.Methods/designStudy designThe TASA Cheque Issue study is an exploratory, parallel-group, unblinded, randomized controlled trial involvingthe allocation of 273 participants to either no intervention(control group), where participants continue to receivetheir provincial social assistance on government chequeissue day, or to one of two intervention groups: (1)monthly social assistance payment on a day that does notfall during cheque issue week (staggered group), or (2)semi-monthly on days that do not fall during cheque issueweek (split and staggered group). Participants will beunder active observation for six income assistance cycles,or approximately six months. The primary outcome ofinterest is increased frequency, street value, or numbersubstances other than cannabis used in the 3 days startingwith government cheque issue day as assessed by the stan-dardized Timeline Followback instrument [30, 31]. Thepre-specified secondary outcomes are: overall monthlydrug use; increased drug use on individual cheque issuedays; drug-related risk, including non-fatal overdose;Richardson et al. BMC Public Health  (2016) 16:668 Page 3 of 16barriers to service access; exposure to violence; interac-tions with police; emergency department (ED), emergencydepartment mental health (EDMH) and substance usehospitalization (SUH) admissions; leaving hospital againstmedical advice (AMA); health care interruption or discon-tinuation; and changes in spending patterns as repre-sented by the amount of time with little or no moneypresent in the participant’s bank account. The TASA studyincorporates a longitudinal nested qualitative parallelprocess evaluation [32] and a cost-effectiveness analysiscomparing the costs of different social assistance disburse-ment arrangements in terms of police, judicial, correc-tions, crime victimization, productivity, and health carecosts. The design of the trial and flow of participants areshown in Fig. 1 (Protocol version 6.6, 14 December 2015).Eligibility and recruitmentIndividuals are eligible for the study if they are 19 yearsof age or older, report active and regular use of drugsother than cannabis, currently receive provincial socialassistance payments on a monthly basis, report intensi-fied drug use at the time of cheque issue days in the sixmonths prior to recruitment, are eligible and willing tobe a client of Pigeon Park Savings, and are not currentlyadministered (where a third party manages their socialassistance, often disbursing funds in smaller amounts).Cannabis is excluded from the TASA outcomes of inter-est and study eligibility criteria due to its lower risk pro-file [33] and to be consistent with standard researchpractice in the study context [34, 35]. Individuals areineligible if they have imminent plans to relocate outsidethe greater Vancouver area or discontinue their socialassistance receipt, have outstanding criminal justice sys-tem involvement that could result in incarceration, orhave been barred from membership at Pigeon Park Sav-ings. Recruitment follows a multi-pronged approach.Three ongoing prospective cohort studies of people whouse illicit drugs in Vancouver, Canada [36] have recentlyadded the following question to their research instru-ments: “Did any of the following things ever happen toyou in the days around cheque day?” The question isfollowed by a set of response options representing differ-ent types of drug-related harm, and is used to identifyprospective participants for referral to the TASA study.Additional recruitment efforts utilize community-basedmethods, including advertisements at PWUD advocacyorganizations, street-based outreach and word of mouththrough the study team’s established contacts. Localhealth, social and financial and service providers haveadditionally agreed to refer clients who meet theFig. 1 Flow of participants through the TASA trialRichardson et al. BMC Public Health  (2016) 16:668 Page 4 of 16eligibility criteria and to post advertisements for thestudy in their offices. The study also recruits participantsthrough the network of addiction physicians practicingin the local health authority. Recruitment takes place ona continuous rolling basis over an 18-month period, withthe final participant completing the intervention 2 yearsafter the start of recruitment.Screening, consent and baseline assessmentIndividuals expressing interest in the study are adminis-tered a brief screening questionnaire by study staff to deter-mine whether they fulfill eligibility criteria. This includesconfirmation that participants are willing and able toadhere to study procedures. Participants are required toprovide proof of non-administered provincial social assist-ance receipt in the form of an official statement of benefits,recent bank statement, or cheque stub from a recent socialassistance cash transfer. Additionally, government-issuedidentification is needed as verification of identity and toopen a bank account at Pigeon Park Savings. Study proce-dures and requirements are then explained to potentialparticipants to confirm participant interest and suitability.During the consent process, research staff explainstudy procedures, potential risks and benefits of partici-pation, planned data linkages to third party administra-tive data sets and other consent information outlined inthe informed consent form. As is common for RCTs,after signing the consent form participants undergo ashort oral “consent quiz” to ensure full comprehensionof key points such as voluntary consent, randomization,confidentiality, study design and procedures, and with-drawal processes. Individuals receive additional informa-tion from study staff on any areas of misunderstandingidentified by the quiz.Following informed consent and prior to randomization,participants complete an interviewer-administered base-line assessment. This assessment includes the collectionof demographic information and completion of question-naire items to assess drug use, income generation andmaterial security, drug-related activities and exposures,criminal activity, police contact, exposure to violence,drug debt, addiction treatment enrolment, health andsocial service use, cheque day activities and health relatedquality of life (see Measures section and Table 1). Baselinemeasures are collected for past 6-month as well as past 2-week time frames to ensure sufficient background infor-mation as well as data consistent with the 2-week timeframe of follow-up observations.Randomization, allocation concealment and blindingOnce participants complete the consent process and base-line procedures, they are randomized to one of the con-trol, staggered, or staggered and split study arms by thestudy coordinator using a 1:1:1 allocation ratio. The studyemploys a stratified block randomization procedure [37]where randomly sized blocks comprised of equal numbersof recipients of the three main types of provincial socialassistance (standard employable income assistance sup-port, persons facing persistent multiple barriers, personswith disability) are allocated to each of the three studyarms. Block sizes are random multiples of three to ensurethe proportional allocation of recipients of each categoryof social assistance to each study arm. The randomizationalgorithm was developed by the study statistician in SASsoftware v9.4 (SAS, Cary, North Carolina, USA) and wasembedded in the participant tracking system to allow foreasy, on-site randomization following the completion ofall study enrollment and baseline procedures andallocation concealment until randomization.Notably, it is not possible to conceal study arm alloca-tion from participants, given the potentially significantchanges to their social assistance arrangements from thestudy intervention. It is also not possible to have anassessor-blinded trial, as study assessors track participantactivities in relation to government as well as individualcheque issue days, thereby revealing study arm alloca-tion information through data collection activities (seeMeasures section).InterventionsControl groupParticipants randomized to this study arm receive nointervention. They continue to receive monthly socialassistance payments according to the British ColumbiaMinistry of Social Development and Social Innovationdisbursement schedule.Staggered social assistance disbursement (staggeredParticipants in the ‘staggered’ intervention study armhave access to their social assistance payments on amonthly basis on a day that does not fall during govern-ment cheque issue week and that is not the same as allother participants in the staggered arm. Due to the vari-ation in government cheque issue day timing, whichranges from the 16th to the 29th of the month, individualcheque issue is allocated relative to government chequeissue day (e.g., Tuesday the week after cheque issue).Participants undergo a transition period during whichthey receive their social assistance up to 7 days latereach month until they reach their study cheque issuedate. This transition period ensures that participants donot have a time period between payments any longerthan the longest period between payments on thegovernment schedule, which is 35 days. To facilitatetimely transitions to the new disbursement schedulewithin the timeframe of the intervention, possibilities forindividual cheque issue is are limited to business daysduring the first and second week following governmentRichardson et al. BMC Public Health  (2016) 16:668 Page 5 of 16Table 1 Measures used in the TASA TrialFollow-up (FU) period Post-follow-upMeasure(s) Screen Baseline FU 1-13 Intervention withdrawal Study exit 60-day visit Data linkagePrimary Outcome MeasureDaily Drug Use (TLFB) ✓ ✓ ✓Demographic MeasuresDate of Birth ✓ ✓Gender ✓Ethnicity/race ✓Relationship status ✓Educational attainment ✓Residency / housing status ✓ ✓ ✓ ✓Drug Related Activity/ExposuresDrug use (past six months) ✓Expenditure on drugs ✓ ✓ ✓Binge drug use ✓ ✓ ✓Distributive and acquisitive syringe sharing ✓ ✓ ✓Crack pipe and drug equipment sharing ✓ ✓ ✓Assistance injecting ✓ ✓ ✓Public drug use ✓ ✓ ✓Non-fatal overdose ✓ ✓ ✓Addiction treatment and harm reductionAddiction treatment (type, timing) ✓ ✓ ✓Treatment interruptions/missed visits ✓ ✓ ✓Supervised injection facility use ✓ ✓ ✓Police contact and illegal activityPolice contact (frequency, type) ✓ ✓ ✓Criminal activity ✓ ✓ ✓Exposure to ViolenceExposure to violence (frequency, type) ✓ ✓ ✓Type of perpetrator ✓ ✓ ✓Police/medical involvement ✓ ✓ ✓Timing ✓ ✓ ✓Health and Social Service UseService accessed ✓ ✓ ✓Barriers to service access (type, timing) ✓ ✓ ✓Missed appointments ✓ ✓ ✓Leaving hospital against medical advice ✓ ✓ ✓Income and Financial InformationSocial assistance income ✓ ✓Additional Income Sources ✓ ✓ ✓Material security ✓ ✓ ✓Daily income generation activity ✓ ✓ ✓Banking practices ✓Drug Debt ✓ ✓ ✓Government cheque day activities ✓ ✓ ✓Richardson et al. BMC Public Health  (2016) 16:668 Page 6 of 16cheque issue day. This allows for 10 possible paymentschedules for participants in the staggered arm. Theexact date of disbursement is determined at randomthrough a second randomization procedure. Study staffprovide each participant with a detailed payment sched-ule to assist participant recall.Staggered and split and social assistance disbursement(staggered and split)Participants in the ‘staggered and split’ interventionstudy arm have their social assistance payments releasedtwice a month on days that do not fall during the weekof government cheque issue. The days of income assist-ance receipt are spaced 2 weeks apart and determined atrandom. As with the staggered intervention arm, indi-vidual cheque issue days are established in relation togovernment cheque issue day and participants transitionto their new schedule incrementally. The first individualcheque issue day falls in the week following governmentcheque issue, with the second occurring on the corre-sponding day of the week 2 weeks later. This could be,for example, Tuesday of the week following governmentcheque issue day and Tuesday 2 weeks after that day.This allows for five randomization possibilities for the stag-gered and split intervention study arm. Participants areagain provided with a schedule of their individual chequeissue days for the duration of their study participation.The change in social assistance payment schedules forboth intervention arms is managed through Pigeon ParkSavings, which has developed a system, in conjunction withstandard social assistance direct deposit services from theBritish Columbia Ministry of Social Development andSocial Innovation, that varies when individuals have accessto their social assistance payments. Social assistance pay-ments are directly deposited into the participant’s account,and the system “locks” their payment so that these fundscannot be withdrawn. Money is “unlocked” according tothe participant’s disbursement schedule that determinesboth the timing and frequency at which payments arereleased. Intervention arm participants can develop person-alized arrangements that ensure access to essential funds atspecific times to prevent housing instability or paymentdefault. A study liaison at Pigeon Park Savings managesaccount creation, direct deposit requests to the Ministryand the implementation of holds and releases of funds. Ac-counts include unlimited withdrawals and a bank card withno-fee withdrawals at credit union automated tellermachines. Pigeon Park Savings banking fees are charged ata rate of $5/month, and individuals are not required to paythis fee while they are enrolled in the study.Follow-up assessmentsOnce individuals enrol (control arm) or receive theirfirst payment on their individual schedule (staggered andTable 1 Measures used in the TASA Trial (Continued)Individual cheque day activities ✓ ✓Health-related quality of lifeEuro-QoL (EQ-5D) ✓ ✓ ✓Study-related measuresMotivation to participate ✓Treatment preferences ✓ ✓Client satisfaction questionnaire (CSQ-4) ✓ ✓Reasons for intervention withdrawal ✓Duration of intervention ✓Participant experiences ✓External Data SourcesHospital, ED, EDMH, SUH recordsCommunity and primary service records ✓Emergency health services records ✓Supervised injection facility records ✓Prescribed medications ✓Banking records (VanCity) ✓Ministry of Social Development & Social ✓Innovation assistance receipt records ✓Police contact records ✓Abbreviations: FU follow-up, TLFB timeline follow back, ED Emergency Department, EDMH Emergency Department Mental Health, SUH substanceuse hospitalizationRichardson et al. BMC Public Health  (2016) 16:668 Page 7 of 16staggered and split) arms, follow-up visits occur every2 weeks for the 26-week period of active study participa-tion, for a total of one baseline and 13 regular follow-upinterviews. Frequent follow-up visits mitigate recall-related threats to data reliability for key measures [38].During study follow-up visits, a trained intervieweradministers the follow-up questionnaire, which includespast 2 week assessments of all study measures. Participantsafety, including the monitoring of adverse events orserious adverse events related and unrelated to study par-ticipation is also assessed at every follow-up. Follow-upvisits take 20–30 min. Due to the intent-to-treat nature ofthe study, participants can withdraw from the interventionbut are encouraged to continue to complete regularfollow-up visits.Qualitative parallel process evaluationThe TASA study also includes a qualitative parallelprocess evaluation concurrent to the quantitativesurvey-based evaluation of the study interventions.This evaluation is comprised of a longitudinal, nestedstudy involving a sub-population of TASA study par-ticipants who complete three qualitative interviewseach, the first at baseline, the second at their studymid-point and the third at study completion or with-drawal. Following randomization to the study inter-ventions, approximately 45–50 participants areselected to be participants in a qualitative parallelprocess evaluation. Quota sampling ensures hetero-geneity across categories of social assistance, drug usepractices and study arms, with a weighted distributionfocusing on the experiences of individuals allocated tothe intervention arms (8–10 participants recruitedfrom control versus 15–20 from each interventionarm). Following a separate informed consent processspecific to qualitative participants, audio-recorded in-terviews are conducted by trained qualitative inter-viewers in private settings at the study site. Datacollection follows principles of data saturation inwhich the study team continues to recruit participantsuntil no new themes emerge in relation to key topicsfrom interviews among individuals in each of thestudy arms [39, 40]. Interviews employ topic guidesinformed by previous studies conducted by the inves-tigative team [5, 6, 41] and a comprehensive reviewof the relevant literature. Baseline interviews exploreprior experiences of social assistance receipt, incomeand material hardship, substance use patterns, and fi-nancial management. Midpoint interviews focus ontransitions to new payment schedules as well aschanges in material circumstances, financial manage-ment, drug use patterns and expenditure on drugs.Exit interviews explore study experiences of socialassistance receipt and broader changes to life circum-stances (e.g. income, housing), drug use and drug-related activity, and reasons for study or interventionwithdrawal, where applicable. Initial interviews informsubsequent interviews among the same participants,building on previous responses and benefits or chal-lenges previously identified by participants.Study completion and post-evaluation questionnairesAt the final regular follow-up study visit, participants areadministered a study completion questionnaire thatassesses, in addition to regular follow-up measures, theirparticipation experiences and opinions about the inter-vention. Sixty days post study completion, a follow-upinterview is administered to assess the safety of partici-pants and any significant changes following the comple-tion of the study protocol. Participants in all study armsare able to access the study intervention of their choicefollowing the completion of their participation shouldthey wish to do so. Screening, baseline, follow-up, quali-tative and post-evaluation study visits take place at apurpose-built field research office located in the DTESand operated by the British Columbia Centre for Excel-lence in HIV/AIDS. If needed, follow-up visits can takeplace at the participant’s residence, over the phone, orother space considered safe by the participant.Participant honorariaConsistent with standard practice in research involvingPWUD [34, 42], participants are compensated for theirtime and interview-related expenses with honoraria in thefollowing Canadian dollar (CAD) amounts: $30 for base-line/randomization visit; $10 per follow-up interview, withfollow-up interview incentive bonuses following the firstpost-baseline follow-up ($10), the completion of fivefollow-up interviews ($15), the completion of nine follow-up interviews ($20) and the completion of the final follow-up interview ($25). The post study follow-up interviewhonorarium is $15. The honorarium for each qualitativeinterview is $30 per interview. Participants will thereforereceive a maximum of $245 for participation in the quanti-tative portion of the study and $335 for participation inboth the qualitative and quantitative study components.MeasuresTable 1 outlines the measures taken at screening, baseline,follow-up, withdrawal and study exit interviews andthrough post-study data linkages. For all primary and sec-ondary outcomes of interest described below, we collectdata on indicators on government cheque issue days bytracking activity and exposures on a daily basis or throughspecific questionnaire items and response options. Wheregovernment cheque issue and individual cheque issuetiming are not the same (i.e., for intervention participants),Richardson et al. BMC Public Health  (2016) 16:668 Page 8 of 16we additionally ask specific questions about individualcheque issue day activities and exposures. By collecting in-formation about the occurrence of key outcomes of inter-est on non-cheque issue days as well as those coincidingwith government and individual payments, we are betterable to distinguish between individual cues for increaseddrug use coinciding with individual payments and socialcues prompted by synchronized payment across the popu-lation on government cheque issue day.Socio-demographic information and housing statusSocio-demographic information, collected at baseline,includes standardized measures of age, gender, ethnicity,relationship status and highest level of educationalattainment. Self-reported measures of residency andhousing status include neighbourhood of residence, typeof residence (e.g. house, apartment, single room occu-pancy hotel [43], no fixed address), and housing stability,measured by the number of residences occupied in thelast six months (baseline) or past 2 weeks (baseline andfollow-up).Drug use (timeline follow back; primary outcome measure)The TASA study’s primary outcome is increased in-tensity of drug use on government cheque issue day(i.e. a binary outcome). Participant drug use is consid-ered to have intensified if, in the 3 days starting withgovernment cheque issue day, a participant increasesby 40 %: (1) the daily frequency of non-cannabis druguse; (2) the street value of drugs consumed; or (3)the number of non-cannabis substances used, includ-ing alcohol and illicit prescription opioids, all com-pared to the average frequency, value of drugs usedor number of substances used on all other days ofthe calendar month. For example, if an individualuses crack cocaine once daily on non-cheque issuedays and does not use methamphetamine, but in-creases uses crack cocaine 3 times per day and meth-amphetamine once per day on the 3 days beginningwith cheque issue day, they will have intensified theiruse according to the first and second criteria. Thesecut points were selected based on prior empiricalstudies [8] and our intention to capture changesamong high-intensity users that could be overlookedif more blunt measures common to RCTs amongPWUD, such as daily drug use [44], are used. Staffgather data measuring the primary outcome using theTimeline Follow Back (TLFB) instrument, a reliable,validated, calendric instrument that collects daily in-formation on drug use patterns, including drugs usedand frequency of use [30, 31]. We additionally collectparticipant estimates of the street value of drugs usedas a proxy for quantity or dosage of drug consumed,which may not be captured by drug use frequencymeasures [45, 46]. We also use daily TLFB drug usedata to assess our key secondary outcome of intensi-fied drug use on individual cheque issue days, identi-fied using the same frequency, street value or numberof drugs criteria as intensified drug use ongovernment cheque issue day.Drug-related activities and exposuresA range of activities and exposures has been associatedwith increased risk of drug-related morbidity and mortalityas well as social and community harms [47–51]. Tosupport analyses of secondary outcomes of interest, past sixmonth (baseline) and past 2 week (baseline and follow-up)self-reported measures for drug use-related activities andexposures include: changes in drug use patterns (e.g., bingeor greater than average drug use); distributive and acquisi-tive syringe, crack pipe or drug equipment sharing; receiv-ing an assisted injection; public injection and non-injectiondrug use; and non-fatal overdose. Questionnaire items havepreviously been verified for use in the current study contextthrough longstanding cohort studies involving PWUD [52].Addiction treatment enrolment and harm reduction serviceusePast six month (baseline) and past 2 week (baseline andfollow-up) data on addiction treatment enrolment arecollected for secondary analyses about addiction treat-ment interruption. Questions include the type of addic-tion treatment, interruptions to ongoing opioid assistedor other treatments, missed visits, the reasons for addic-tion treatment interruption (if any), and the impacts oftreatment interruptions or missed visits (e.g., relapse,withdrawal). We additionally ask participants about pastsix month (baseline) and past 2 weeks (baseline andfollow-up) use of Insite, a local supervised injection facil-ity (SIF), given the association between exposure to theSIF and significant improvements in in morbidity, mor-tality, uptake of addiction treatment uptake and mea-sures of public order [53].Police contact and illegal activityWhile drug market enforcement has been shown to haveadverse public health and social impacts [54], studieshave also identified police as critical supports for publichealth and community safety initiatives [55]. We collectdata on all police interactions, and the measures used inthe current study facilitate distinctions between differenttypes of police contact (e.g., service referral, arrest), aswell as participant perceptions of the reason for contact.In support of economic analyses, we also collect dataabout engagement in illegal activity to assess any changesin the costs of criminality and victimization as a result ofRichardson et al. BMC Public Health  (2016) 16:668 Page 9 of 16the intervention, with past six month (baseline) and past2 week (baseline and follow-up) recall periods.Exposure to violenceConsistent with previous analyses documenting the rela-tionship between socio-economic marginalization andexposure to violence [56], the TASA study also seeks todocument whether varying social assistance paymentsimpacts participants’ exposure to violence. The study in-strument includes measures to identify the frequency,timing and type of violence victims were exposed to, thetype of perpetrator (e.g. stranger, intimate partner) andwhether the participant sought police or medical atten-tion. The study solicits analogous information regardingthe perpetration of violence.Health service utilizationSeveral key secondary outcomes of the TASA study arelinked to whether the study intervention can reduce orsmooth health service utilization. Previously identifiedhealth service impacts of synchronized social assistance in-clude ED, EDMH or SUH admissions [7, 11–13, 19, 57];discharges against medical advice directly prior to chequeissue [7, 9, 12]; health service access barriers linked to ele-vated demand [6]; and interruptions to ongoing medicalcare [9, 22]. In support of these assessments the studyinstrument collects past six months (baseline) and past2 weeks (baseline and follow-up) information on whichhealth services were accessed, whether participantsattempted but were unable to access a service, the tim-ing and reason for encountering service access barriersand whether participants missed medical appointmentsaround cheque issue. We additionally collect informa-tion on whether, if hospitalized, participants left hos-pital against medical advice, as well as the reasons forand timing of their departure.Social service utilizationSocial assistance recipients’ use of social services, suchas meal programs, outreach services or drop-in centres,may similarly undergo cyclical changes or face barriersto accessing services at times when these are oversub-scribed. Similar to health care utilization indicators, thestudy instrument includes questions asking participantsabout social service access, barriers to accessing servicesand the timing of service access barriers in relation tocheque issue in the past six months (baseline) and past2 weeks (baseline and follow-up).Income and material securityConsideration of income generation is critical given thatmany social assistance recipients supplement their incomewith illegal or prohibited income generation (e.g., sexwork, drug dealing) [14, 15], and that such activities areassociated with health harms [14, 15, 56]. Therefore, chan-ging assistance patterns may change patterns of other in-come generation. To assess the impact of the interventionon financial well being, we measure income generation ac-tivity, income amounts and material security in the pastsix months (baseline), and past 2 weeks (baseline andfollow-up). Data referring to the past six months is col-lected using an instrument verified for use in the currentstudy context [15, 58]. Data referring to the 2 weeks priorto interview is collected by expanding the use of theTimeline Followback instrument to track daily incomegeneration information, including sources and estimateddaily earnings. Material security in the past six months(baseline) and past 2 weeks (baseline and follow-up) isassessed using a validated scale [59] that has been adaptedto the current study context. Further measures includeinvolvement in acquisitive criminal activity, such as theft,as well as amounts of drug debt accumulation and towhom drug debt is owed.Quality of lifeTo support health economic evaluation, the CanadianEuroqol Group EQ-5D [60, 61] will be administered atbaseline and all follow-up visits. The EQ-5D can beadministered quickly and measures health related qualityof life and includes participant self-reports of mobility,self-care, usual activities, pain/discomfort and anxiety/depression, as well as a health thermometer ranking of aparticipant’s overall state of health, and has beenvalidated for use among PWUD [62].Study-related measuresTo assess considerations related to the study and studyintervention that may affect study outcomes, at baselinewe collect information on individual’s motivations toparticipate in the study (e.g., financial compensation,desire to reduce drug use around cheque issue) and theirtreatment arm preference. At baseline and the finalfollow-up visit, participants complete the client satisfac-tion questionnaire (CSQ), a validated instrument [63]that assesses satisfaction with services and is used hereto measure participant perceptions about social assistancereceipt, including the degree to which social assistance andrelated Ministry of Social Development and SocialInnovation supports meet needs, help with problems, andare satisfactory. Additionally, for participants that withdrawfrom the intervention prior to completing their 26-weekintervention period, questions are asked about theirreasons for and circumstances surrounding withdrawal. Atthe final study follow-up we assess participant experiences,including participant opinions about the benefits anddrawbacks of their intervention and best options forcheque issue timing and frequency.Richardson et al. BMC Public Health  (2016) 16:668 Page 10 of 16Participant safetyThe design of the current trial does not involve anychange in the amount of social assistance provided toparticipants, and we therefore do not anticipate anynegative changes to their financial well-being. However,desynchronizing social assistance may disrupt moneymanagement strategies, potentially interfering with thepayment of expenses. Importantly, most participants inthe TASA study have the rent portion of their socialassistance paid directly to their landlords. Participantswho have rent or other expenses can make arrangementson a case-by-case basis with Pigeon Park Savings toensure the availability of funds at the appropriate time.Additionally, for staggered and split arm participants,receiving multiple payments per month may result inmore frequent periods of intensified drug use, albeitthese periods will likely be shorter or less intense giventhat less money will be disbursed at any one time. Forintervention participants, there may also be challengestransitioning onto or off of a new social assistancepayment schedule at the start or end of the study period.To mitigate this concern, study staff members canprovide referrals to financial management supportservices as appropriate. We do not anticipate the studywill expose participants to unusual risks. Nevertheless,we monitor participant safety as a part of each bi-weeklyparticipant assessment, including considerations of foodand housing insecurity, exposure to violence, access tocritical health and social services and other emergingissues. We follow standard reporting procedures for ad-verse events and serious adverse events [64], includingoversight from an external, independent Data Safety andMonitoring Committee (DSMC). Participants are able towithdraw from the intervention, the study, or both atany point, and may be withdrawn from either by investi-gators in the event of concerns over safety.Sample sizeWe base sample size calculations for the TASA study onan a priori minimum clinically important difference(MCID) in the rate of individuals reporting intensifieddrug use on government cheque issue days between con-trol and non-control participants. For the purposes ofsample size calculations, comparisons were drawn be-tween control and staggered arm participants. We an-ticipate this intervention will have a lesser impact on theprimary outcome of interest than the staggered and splitarm and should therefore provide basis for sample sizecalculations. Given a lack of prior comparable experi-mental research, the MCID was set conservatively to a20 % difference between control and staggered interven-tion arm participants. This MCID was based on resultsfrom previous observational studies that report increasesin the rates of drug related harm between 24.9 and81.2 % in the days surrounding government cheque issue[5, 7, 12, 13, 17–20, 57]. We combined this MCID withan estimate of 85 % of individuals in the control groupreporting intensified substance use on communitycheque issue day (Pc). This was set lower than 100 % toaccount for potential reporting biases at study screening,where all participants must indicate elevated use inorder to be eligible, as well as any study participationeffect on the outcome of interest among control armparticipants.Sample size calculations used methods to detect differ-ences between proportions in a repeated measures de-sign and were calculated using Power and Sample Size(PASS v.12) software. Sixty-five participants per armcompleting all follow-ups would allow the detection of a20 % difference in the rates of intensified drug usebetween control and staggered intervention arm partici-pants with 80 % power at a 5 % significance level. Thesecalculations assumed that control and intervention armrates of intensified drug use follow a Bernoulli distribution,the number of government cheque issue days fallingwithin the observation period is 6 and, as in previous stud-ies involving PWUD [65], the autocorrelation betweenmeasures of the same individual over time (ρ) is 0.6. Basedon previous studies among PWUD, and noting the import-ance of accounting for missed follow-up visits and loss tofollow-up [66], we additionally account for observation-level non-response of 33 % and a potential rate of loss tofollow-up of 25 %, reaching a final sample size of 91individuals per arm or 273 in total across all three arms.The possibility of type II, false negative findings, could re-sult from insufficient power, or, in the event of interven-tion withdrawal, the dilution of the effect of theintervention in ITT analyses from participants who haveresumed regular synchronized payments but are analyzedas part of the intervention arm. These risks will be partiallymitigated by the inclusion of sensitivity analyses, as out-lined in the Data Analysis section below.Data linkagesTo verify participant self-report wherever possible, andsupplement data for the assessment of cheque issue-related costs incurred by institutions that provide social,health, financial and security (i.e., policing) services toPWUD and their communities, we will obtain third partyservice provision and administrative records from externaldatabases on all participants. Participant records will berequested from hospitals, health service providers, theprovincial prescription medication database, ProvincialMinistry of Social Development and Social Innovation,study partner credit union, and police and criminal justicedatabases. Records will be requested for a 30-monthperiod beginning 1 year before the beginning of activestudy participation through to 1 year after following theRichardson et al. BMC Public Health  (2016) 16:668 Page 11 of 16completion of active study participation in order to assessthe impact of study involvement on access to medical andsocial services over time. Data linkages will be establishedthrough participant names, birth dates and participants’personal health number, a unique and persistent identifierissued for medical billing and tracking purposes to all resi-dents of British Columbia. Data will be de-identified toprotect participant confidentiality.Data analysisQuantitative analysesAt the conclusion of the trial, the primary and secondaryendpoints will be assessed via intention-to-treat analyses[67], including all patients randomized regardless of com-pliance with the intervention protocol. Sub-analyses willinclude a modified intention-to-treat analysis that will useall participants with at least one follow-up [67, 68]. First,we will compile descriptive characteristics of the sample,assessing for systematic differences in key characteristicsacross study arms using the Chi-Square test for binary vari-ables and the Mann-Whitney U test for continuous vari-ables. The primary outcome of interest will be derivedfrom TLFB information as a repeated binary variable (in-tensified drug use on the 3 days beginning with govern-ment cheque issue day vs. not) and will be quantified asthe proportion of individuals TASA participants reportingincreased drug use during this period at each governmentcheque issue. This outcome, and other secondary out-comes similarly quantified will be examined using pooledgeneralized linear mixed effects models, with binomial dis-tribution and logit link specified to produce unadjustedand adjusted odds ratios and 95 % confidence intervals.Other secondary outcomes of interest, quantified as countsof the number of occurrences over the course of the obser-vation period (e.g., hospital ED admissions), will be exam-ined using generalized linear mixed-effects model, with aPoisson distribution and a log link specified for count datawith corrections for overdispersion, including a Vuong testfor overdispersion and, where indicated, the use of a nega-tive binomial model [69]. All analyses will include binaryindicators of study arm allocation to assess for systematicdifferences in primary and secondary outcomes of interestacross study arms. Multivariate analyses will be adjustedfor: (1) binary indicators of the type of income assistancethe participant receives; (2) an indicator variable forwhether the individual changed their financial service pro-vider to Pigeon Park Savings (PPS) at the start of the studyto control for the effect of becoming a PPS client; (3)socio-demographic characteristics, including gender, ethni-city and age; indicators of addiction treatment enrollmentand DTES residency, and (4) a categorical variable for thenumber of days between government cheque issue andtheir study visit to account for differences in recall reliabil-ity. As is common in RCT analyses [70, 71], missingassessments of intensified drug use will be treated aspositive and tested for sensitivity to this assumption[72, 73]. Additional sensitivity analyses will include: (1)a modified intention-to-treat analysis using all partici-pants with at least one follow-up assessment [74]; and(2) per-protocol analyses of participants who maintainenrollment in the intervention [75, 76].Qualitative analysisFollowing verbatim transcription of interview audio-recordings, we will inductively generate a coding frame-work capturing a priori key analytic constructs derivedfrom the topic guide as well as emergent themes derivedfrom interview transcripts, continuously refining andconsolidating code categories. Qualitative data will becontextualized through linkages to TASA questionnairedata in terms of individual participation trajectories,adverse events and serious adverse events. Data fromsecond and third interviews will test the boundaries ofexisting categories where overlap in substantive contentexists, identify new conceptual categories and examinenegative or inconsistent evidence. Particular focus willbe on experiences of synchronized and unsynchro-nized cheque issue days among intervention partici-pants to explore the role of individual and social cuesfor drug use. We will also compare experiences ofdrug use, drug-related harm and changes to financialmanagement strategies between individuals allocatedto different intervention arms. Longitudinal data col-lection will provide opportunities for participants toprovide feedback on emerging analyses and reflect onearlier study experiences.Economic evaluationThe economic evaluation of the TASA study will involvea cost-utility analysis to determine the cost-effectivenessof the study interventions. We will adopt a societal per-spective that includes the direct (healthcare utilization:medication, inpatient and outpatient care) and indirect(criminal activity and lost productivity) consequences ofdifferent social assistance disbursement arrangements.Incremental cost-effectiveness ratios (ICERs) will be calcu-lated for the 26-week timeframe of individual using trial-based analysis, contrasted with results from a model-basedanalysis projecting up to a lifetime horizon by adapting apreviously-developed model [77]. Statistical analyses willinclude: (1) Cost estimation; (2) Health-related quality oflife (HRQoL) assessment and valuation; (3) Incrementalcost-effectiveness analysis; and (4) Analysis of uncertainty.The relevant costing data includes intervention costs,health resource utilization and criminality, all collected atfollow-up via standardized or verified instruments, withpreviously-used unit cost sources [77]. The number ofquality-adjusted life-years (QALYs) [78] accumulatedRichardson et al. BMC Public Health  (2016) 16:668 Page 12 of 16during the study period for each individual will be esti-mated using the Euroqol EQ-5D [62]. We will augmentstandardized measurement by quantifying QALY loss dueto overdose via direct assessment at baseline.ICERs will be calculated in the typical formulation,for both the trial- and model-based analyses [79].Additionally, we propose a net-benefit regression ap-proach for the trial-based analysis, using establishedeconometric techniques [80], allowing for subgroupassessment of cost-effectiveness. Uncertainty aroundICERs will be quantified using non-parametric boot-strapping and Monte Carlo simulation for trial- andmodel-based analyses, respectively [81]. This informationcan then be used to inform whether it is worthwhile tocollect further information to clarify the decision rule, inthe event that there is no clear choice among control orintervention strategies, or if there is a high level of uncer-tainty in the decision [82]. The analysis will conform toguidelines on cost-effectiveness analyses conducted along-side clinical trials [77, 83].Study oversightTASA is monitored by an independent DSMC comprisedof a biostatistician, epidemiologist, addictions physicianand community service provider with detailed knowledgeof the interests of the target study population. The DSMCcan recommend study stoppage for reasons of efficacy, re-cruitment or participant safety, and meets semi-annually.Data concerning adverse events and serious adverseevents are provided in real time, and study investigatorssubmit quarterly reports on participant recruitment andretention. Interim analyses whose results are to be re-ported to the DSMC are scheduled following the recruit-ment and study completion of the first third and secondthird of study participants. Any protocol modificationswill be subject to further approval form the ethical reviewand reported to all study investigators, trial registries,journals and, where relevant, trial participants.Data management and quality assuranceEach participant is assigned a unique numeric study iden-tifier code at the beginning of their enrollment in theTASA trial to enable the de-identification of data. Confi-dential electronic questionnaire data and completed hardcopy forms will be entered into a password-protectedOracle database. Qualitative data stored in electronic files,including interview recordings, interview transcripts aswell as field-notes will be password protected. Hard copiesof field notes, consent forms, and interview transcripts willbe stored in locked filing cabinets in the secure fieldresearch office in space restricted exclusively to study staffand maintained by study investigators. All participant dataand digital research documents are stored in encryptedfiles on secure servers at the British Columbia Centre forExcellence in HIV/AIDS, and only the Principal Investiga-tor and key study personnel have access to any datacontaining personal identifiers to facilitate participantsafety and monitoring and data linkages. Study personnelperform random checks to verify the accuracy of data in-put into the clinical trial database, and triangulation willbe undertaken to identify and rectify any deficiencies toensure data integrity. Trained personnel conduct theTASA study according to standard operating proceduresand the principles of Good Clinical Practice.DiscussionThe unintended health, social and economic harms result-ing from synchronized monthly social assistance are welldocumented [5–7, 9–13, 16–24]. However, the lack of con-trolled experimental studies examining alternate disburse-ment strategies impedes the development of evidence-informed policies that could preserve the important finan-cial security and other benefits of social assistance while de-creasing their role in the production of drug-related harm.The TASA study is the first RCT that begins to address thesignificant gap in our understanding of potential alterna-tives to a considerable driver of avoidable morbidity andmortality, elevated costs to communities and challenges toservice provision. By providing robust research exploringinterventions to address both the individual and social ef-fects of current disbursement policies, the study is beingconducted with a view to identify strategies that are betterable to promote public health and safety. This study there-fore has considerable implications for policy makers giventhe longstanding and widespread recognition ofdisbursement-related harm in multiple jurisdictions [7, 8,21, 84].The study also represents a notable innovation inexperimental and trial-based research in addictions byfocusing on a structural intervention that alters thecharacteristics of an upstream determinant of health-individual income-to mitigate the effects of problematicdrug use and related harms. RCTs among PWUD haveoverwhelmingly assessed behavioural drug use interven-tions, and have, as a result, produced mixed evidence forinterventions that may be difficult to scale up or thatproduce only short-term benefits [85–87]. In contrast,the TASA study examines an intervention with signifi-cant policy relevance and a high degree of scalability,with the possibility for long-term, low-cost or cost-averting implementation. Additionally, the measurementof daily drug use patterns throughout the 26-week dur-ation of each participant’s active study participation pro-duces a robust platform for analyses. The supplementationof study data with information from third-party databaseswill additionally allow for data triangulation and robust ser-vice utilization and cost-benefit analyses to determinewhether either of the tested interventions holds theRichardson et al. BMC Public Health  (2016) 16:668 Page 13 of 16potential to produce system-wide efficiencies and reduceservice provision challenges around government chequeissue. The study additionally benefits from widespreadcommunity acknowledgement of the issues and challengesof synchronized social assistance and strong communitypartnerships in support of the study and the assessment ofalternative approaches.The TASA study uniquely examines an upstreamdeterminant of health and the unintended but negativeconsequences of public policy on the outcomes of vul-nerable and marginalized populations. In testing changesto the structural conditions of social assistance receiptfor PWUD through a community-based RCT, the TASAstudy advances the field of experimental addictions re-search. The inclusion of a paired nested qualitative parallelprocess evaluation and economic cost-effective analysesprovide a robust platform for understanding the impacts,mechanisms of action and cost implications of alternativeapproaches to synchronized monthly social assistancedisbursement. There is significant potential, therefore, forTASA study findings to directly support evidence-informedchanges to social assistance disbursement policy in BritishColumbia, Canada and internationally.AbbreviationsAMA, against medical advice, CAD, Canadian dollars, CRISM, Canadian ResearchInitiative in Substance Misuse, CSQ, client satisfaction questionnaire, DSMC, DataSafety and Monitoring Committee, DTES, Downtown Eastside, ED, emergencydepartment, EDMH, emergency department metnal health, HRQoL, health-related quality of life, ICER, Incremental cost-effectiveness ratios, PPS, PigeonPark Savings, PWUD, people who use illicit drugs, QALY, quality adjusted lifeyearsRCT, Randomized controlled trial, SIF, supervised injection facility, SUH,substance use hospitalization TLFB, Timeline follow back instrumentAcknowledgmentsPigeon Park Savings, a PHS Community Services Society operated branch ofVancouver City Savings and Credit Union, administers the intervention. PPS isnot involved in participant recruitment, randomization or the collection ofbaseline or follow-up data.We thank the following for their contribution to the study: Michelle Davey(Vancouver Police Department), Kenneth Tupper (BC Ministry of Health),Vancouver Area Network of Drug Users (VANDU), Western Aboriginal HarmReduction Society (WAHRS), Joel Singer, Emanuel Krebs, Jennifer Matthews,Tricia Collingham, Kristie Starr, and Ana Prado.LR and MJM are supported by Canadian Institutes for Health Research NewInvestigator Awards and Michael Smith Foundation for Health Research(MSFHR) Scholar Awards. MJM is additionally supported in part by the UnitedStates National Institutes of Health (R01-DA0251525). BM is supported by theNIH (DP2-DA040236) and by a Henry Merrit Wriston Fellowship from BrownUniversity. WS is supported by a MSFHR Scholar Award. EW is supported bya Tier 1 Canada Research Chair in Inner City Medicine. JM is supported bythe British Columbia Ministry of Health and through an Avant-Garde Award(No. 1DP1DA026182) from the National Institute of Drug Abuse (NIDA), atthe US National Institutes of Health (NIH).FundingThe TASA study is supported by the Canadian Institutes of Health Research(MOP 136827, MOP 137068), the Peter Wall Institute, the Michael SmithFoundation for Health Research and a PHCRI and VCHRI joint Innovationand Translational Award funded by the Providence Health Care ResearchInstitute. Funding bodies had no role in the design of the study, the writingof the manuscript or the decision to submit the manuscript for publication.Availability of data and materialUpon completion of data collection, study data will not be publicly availabledue to the sensitivity of data and requirements of third party databaseadministration.Authors’ contributionsLR, MJM and TK conceptualized the study. LR developed the protocol, isthe Principal Investigator on the grants that support the project, oversees allaspects of the study, and wrote the first draft of the manuscript. AL led thedevelopment of operational procedures. All other investigators contributedto the study design and have provided key substantive and editorial inputinto the protocol and manuscript. All authors have read and approved thefinal manuscript.Competing interestsMJSM’s institution has received an unstructured gift from NG Biomed, Ltd.to support his research. JM’s TasP research, paid to institution, has receivedsupport from the BC-Ministry of Health, US NIH (NIDA R01DA036307),UNAIDS, and MAC AIDS Fund. Institutional grants have been provided toJM by Abbvie, BMS, Gilead Sciences, J&J, Merck and ViiV Healthcare.Consent for publicationNot applicable.Ethics approval and consent to participateThis study has received ethics approval from the University of BritishColumbia/Providence Health Care Research Ethics Board (H14-02401). Allparticipants are required to provide written informed consent following priorto their enrollment in the study. Separate release of information consentforms are completed by participants to grant permission to obtain theirpersonal records from external databases, including hospitals, health serviceproviders, the provincial prescription medication database, Provincial Ministryof Social Development and Social Innovation, the study partner credit union,police and criminal justice databases and emergency health services.Author details1British Columbia Centre for Excellence in HIV/AIDS, St. Paul’s Hospital,608-1081 Burrard Street, Vancouver V6Z 1Y6, BC, Canada. 2Department ofSociology, University of British Columbia, 6303 NW Marine Drive, VancouverV6T 1Z1, BC, Canada. 3Faculty of Medicine, Division of AIDS, University ofBritish Columbia, St. Paul’s Hospital, 608-1081 Burrard Street, Vancouver V6Z1Y6, BC, Canada. 4PHS Community Services Society, 20 Hastings Street W,Vancouver V6B 1G6, BC, Canada. 5Faculty of Health Sciences, Simon FraserUniversity, 8888 University Drive, Burnaby V5A1S6, BC, Canada. 6Departmentof Epidemiology, School of Public Health, Brown University, 121 South MainStreet, Providence 02912, RI, USA. 7Department of Emergency Medicine,Faculty of Medicine, University of British Columbia, 910 West 10th Ave,Vancouver V5Z 1 M9, BC, Canada. 8Vancouver Coastal Health, 601 WestBroadway, Vancouver V5Z 4C2, BC, Canada. 9School of Population and PublicHealth, University of British Columbia, 2206 East Mall, Vancouver V6T 1Z3, BC,Canada.Received: 7 July 2016 Accepted: 14 July 2016References1. Boychuk GW. Patchworks of Purpose: The Development of Provincial SocialAssistance Regimes in Canada. Montreal: McGill-Queens University Press;1999.2. Nelson K. Mechanisms of poverty alleviation: anti-poverty effects ofnon-means-tested and means-tested benefits in five welfare states.J Eur Soc Policy. 2004;14:371–90.3. Stephens M. 3rd of tha month”: Do social security recipients smoothconsumption between checks? Am Econ Rev. 2003;93:406–22.4. Epstein DH, Willner-Reid J, Vahabzadeh M, Mezghanni M, Lin JL, Preston KL.Real-time electronic diary reports of cue exposure and mood in the hoursbefore cocaine and heroin craving and use. Arch Gen Psychiatry.2009;66:88–94.5. Zlotorzynska M, Milloy M, Richardson L, Montaner J, Wood E, Kerr T.Timing of social assistance payment and overdose patterns at a Canadiansupervised injection facility. Int J Drug Pol. 2014;25:736–9.Richardson et al. BMC Public Health  (2016) 16:668 Page 14 of 166. Small W, Shoveller J, Moore D, Tyndall M, Wood E, Kerr T. Injection drugusers’ access to a supervised injection facility in Vancouver, Canada: theinfluence of operating policies and local drug culture. Qual Health Res.2011;21:743–56.7. Verheul G, Singer SM, Christenson JM. Mortality and morbidity associatedwith the distribution of monthly welfare payments. Acad Emerg Med.1997;4:118–23.8. Rosen MI. The ‘check effect’ reconsidered. Addiction. 2011;106:1071–7.9. Chan AC, Palepu A, Guh DP, Sun H, Schechter MT, O’Shaughnessy MV, AnisAH. HIV-positive injection drug users who leave the hospital against medicaladvice: the mitigating role of methadone and social support. J AcquirImmune Defic Syndr. 2004;35:56–9.10. Shaner A, Eckman TA, Roberts LJ, Wilkins JN, Tucker DE, Tsuang JW, Mintz J.Disability income, cocaine use, and repeated hospitalization amongschizophrenic cocaine abusers—a government-sponsored revolving door?N Engl J Med. 1995;333:777–83.11. Brunette DD, Kominsky J, Ruiz E. Correlation of emergency health care use,911 volume, and jail activity with welfare check distribution. Ann EmergMed. 1991;20:739–42.12. Riddell C, Riddell R. Welfare checks, drug consumption, and health evidencefrom Vancouver injection drug users. J Hum Resour. 2006;41:138–61.13. Catalano R, McConnell W. Psychiatric emergencies: the check effectrevisited. J Health Soc Behav. 1999;40:79.14. Richardson L, Milloy M, Kerr T, Guillemi S, Hogg R, Harrigan R,Montaner J, Wood E. Socio-economic marginalization and virologicsuppression among antiretroviral therapy-exposed individuals who useillicit drugs. AIDS. 2015;29:2487–95.15. Debeck K, Shannon K, Wood E, Li K, Montaner J, Kerr T. Income generatingactivities of people who inject drugs. Drug Alcohol Depend. 2007;91:50–6.16. Dobkin C, Puller SL. The effects of government transfers on monthly cyclesin drug abuse, hospitalization and mortality. J Pub Econ. 2007;91:2137–57.17. Li X, Sun H, Marsh DC, Anis AH. Impact of welfare cheque issue days on aservice for those intoxicated in public. Harm Reduct J. 2007;4:1–4.18. Catalano R, McConnell W, Forster P, Mcfareland B, Shumway M, Thornton D.Does the Disbursement of Income Increase Psychiatric EmergenciesInvolving Drugs and Alcohol? Health Serv Res. 2000;35:813–23.19. Halpern SD, Mechem CC. Declining rate of substance abuse throughoutthe month. Am J Med. 2001;110:347–51.20. Phillips DP, Christenfeld N, Ryan NM. An increase in the number of deathsin the united states in the first week of the month—an association withsubstance abuse and other causes of death. N Engl J Med. 1999;341:93–8.21. Otterstatter MC, Amlani A, Guan H, Richardson L, Buxton J. Illicit drugoverdose deaths resulting from income assistance payments: Analysis of the‘check effect’ using daily mortality data. Int J Drug Pol. 2016;33: 83–7. InPress.22. Anis AH, Sun H, Guh DP, Palepu A, Schechter MT, O’Shaughnessy MV.Leaving hospital against medical advice among HIV-positive patients. CMAJ.2002;167:633–7.23. Svikis DS, Pickens RW, Schweitzer W, Johnson E, Haug N. Weekly patterns ofdrug treatment attendance. Am J Public Health. 1999;89:752–5.24. Pickett T, Stenstrom R, Abu-Laban R. Association between mental healthapprehensions by police and monthly income assistance (welfare)payments. Can J Psychiatry. 2015;60:146–50.25. Parsons CA, Van Wesep ED. The timing of pay. J Financ Econ.k2013;109:373–97.26. Anakwenze U, Zuberi D. Mental health and poverty in the inner city. HealthSoc Work. 2013;38:147–57.27. Liu S, Blomley N. Making news and making space: framing Vancouver’sdowntown eastside. Can Geogr. 2013;57:119–32.28. Krebs E, Wang L, Olding M, Hayashi K, Milloy M, DeBeck K, Wood E, KerrT, Nosyk B. Increased drug use and the timing of social assistance receiptamong people who use illicit drugs. Under Review.29. Blankenship KM, Friedman SR, Dworkin S, Mantell JE. Structuralinterventions: concepts, challenges and opportunities for research.J Urban Health. 2006;83:59–72.30. Robinson SM, Sobell LC, Sobell MB, Leo GI. Reliability of the timelinefollowback for cocaine, cannabis, and cigarette use. Psychol Addict Behav.2014;28:154–62.31. Hjorthøj CR, Hjorthøj AR, Nordentoft M. Validity of timeline follow-back forself-reported use of cannabis and other illicit substances—systematic reviewand meta-analysis. Addict Behav. 2012;37:225–33.32. Oakley A, Strange V, Bonell C, Allen E, Stephenson J, RIPPLE Study Team.Process evaluation in randomised controlled trials of complex interventions.BMJ. 2006;332:413–6.33. Nutt DJ, King LA, Phillips LD, Independent Scientific Committee OnDrugs. Drug harms in the UK: a multicriteria decision analysis. Lancet.2010;376:1558–65.34. Strathdee SA, Patrick DM, Currie SL, Cornelisse PGA, Rekart ML,Montaner JSG, Schechter MT, O'Shaughnessy MV. Needle exchange isnot enough: lessons from the Vancouver injecting drug use study.AIDS. 1997;11:F59–65.35. Milloy MJ, Kerr T, Bangsberg DR, Buxton J, Parashar S, Guillemi S, MontanerJ, Wood E. Homelessness as a structural barrier to effective antiretroviraltherapy among HIV-seropositive illicit drug users in a Canadian setting.AIDS Patient Care STDS. 2012;26:60–7.36. Urban Health Research Initiative. Drug Situation in Vancouver, 2nd Ed.Urban Health Research Initiative, British Columbia Centre for Excellence inHIV/AIDS, June 2013.37. Efird J. Blocked randomization with randomly selected block sizes. IntJ Environ Res Public Health. 2011;8:15–20.38. Buu A, Li R, Walton MA, Yang H, Zimmerman MA, Cunningham RM.Changes in Substance Use-Related Health Risk Behaviors on the TimelineFollow-Back Interview as a Function of Length of Recall Period. Subst UseMisuse 201439. Patton MQ. Qualitative evaluation and research methods. Newbery Park:Sage; 1990.40. Teddlie C, Yu F. Mixed methods sampling: a typology with examples.J Mixed Methods Res. 2007;1:77–100.41. Kerr T, Kimber J, Rhodes T. Drug use settings: an emerging focus forresearch and intervention. Int J Drug Policy. 2007;18:1–4.42. Davidson P, Page K. Research participation as work: comparing theperspectives of researchers and economically marginalized populations.Am J Public Health. 2012;102:1254–9.43. Shannon K, Ishida T, Lai C, Tyndall MW. The impact of unregulated singleroom occupancy hotels on the health status of illicit drug users inVancouver. Int J Drug Pol. 2006;17:107–14.44. Magill M, Ray LA. Cognitive-behavioral treatment with adult alcohol andillicit drug users: a meta-analysis of randomized controlled trials. J StudAlcohol Drugs. 2009;70:516–27.45. Kilmer B, Everingham SS, Caulkins JP, Midgette G, Pacula RL, Reuter PH,Burns RM, Han B, Lundberg R, Corporation R. What Americas users spendon illegal drugs, 2000–2010. Washington: Office of National Drug ControlPolicy; 2014.46. Singleton N, Murray R, Tinsley L. Measuring different aspects of problemdrug use: methodological developments. 2006. Home Office London.https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/116642/hoor1606.pdf. Accessed 14 February 2016.47. Fairbairn N, Small W, Van Borek N, Wood E, Kerr T. Social structural factorsthat shape assisted injecting practices among injection drug users inVancouver. Can: qual study Harm Reduct J. 2010;7:20.48. Kerr T, Small W, Buchner C, Zhang R, Li K, Montaner J, Wood E. Syringesharing and HIV incidence among injection drug users and increasedaccess to sterile syringes. Am J Public Health. 2010;100:1449–53.49. Miller CL, Kerr T, Frankish JC, Spittal PM, Li K, Schechter MT, Wood E. Bingedrug use independently predicts HIV seroconversion among injectiondrug users: implications for public health strategies. Subst Use Misuse.2006;41:199–210.50. Ivsins A, Roth E, Benoit C, Fischer B. Crack pipe sharing in context: Howsociostructural factors shape risk practices among noninjection drug users.Contemp Drug Problems. 2013;40:481–503.51. Small W, Rhodes T, Wood E, Kerr T. Public injection settings inVancouver: physical environment, social context and risk. Int J DrugPolicy. 2007;18:27–36.52. Kerr T, Small W, Johnston C, Li K, Montaner JS, Wood E. Characteristics ofinjection drug users who participate in drug dealing: implications for drugpolicy. J Psychoactive Drugs. 2008;40:147–52.53. Potier C, Laprévote V, Dubois-Arber F, Cottencin O, Rolland B. Supervisedinjection services: what has been demonstrated? a systematic literaturereview. Drug Alcohol Depend. 2014;145:48–68.54. Werb D, Rowell G, Guyatt G, Kerr T, Montaner J, Wood E. Effect ofdrug law enforcement on drug market violence: a systematic review.Int J Drug Policy. 2011;22:87–94.Richardson et al. BMC Public Health  (2016) 16:668 Page 15 of 1655. Debeck K, Wood E, Zhang R, Tyndall M, Montaner J, Kerr T. Police andpublic health partnerships: evidence from the evaluation of Vancouver’ssupervised injection facility. Subst Abuse Treat Prev Policy. 2008;3:1–5.56. Richardson L, Long C, Nguyen P, Debeck K, Milloy MJ, Wood E, Kerr T.Socio-economic marginalization in the structural production of vulnerabilityto violence among people who use illicit drugs. J Epidemiol CommunityHealth. 2015;69:686–92.57. Maynard C, Cox GB. Association between week of the month andhospitalization for substance abuse. Psychiatr Serv. 2000;51:31.58. Long C, Debeck K, Feng C, Montaner J, Wood E, Kerr T. Income level anddrug related harm among people who use injection drugs in a Canadiansetting. Int J Drug Pol. 2014;25:458–64.59. Ompad DC, Nandi V, Cerdá M, Crawford N, Galea S, Vlahov D. Beyondincome: material resources among drug users in economically-disadvantaged New York city neighborhoods. Drug Alcohol Depend.2012;120:127–34.60. Bansback N, Tsuchiya A, Brazier J, Anis A. Canadian valuation of EQ-5Dhealth states: preliminary value set and considerations for future valuationstudies. PLoS One. 2012;7, e31115.61. Herdman M, Gudex C, Lloyd A, Janssen M, Kind P, Parkin D, Bonsel G, BadiaX. Development and preliminary testing of the new five-level version ofEQ-5D (EQ-5D-5 L). Qual Life Res. 2011;20:1727–36.62. Nosyk B, Sun H, Guh DP, Oviedo-Joekes E, Marsh DC, Brissette S, SchechterMT, Anis AH. The quality of eight health status measures were comparedfor chronic opioid dependence. J Clin Epidemiol. 2010;63:1132–44.63. Attkisson CC, Zwick R. The client satisfaction questionnaire. Psychometricproperties and correlations with service utilization and psychotherapyoutcome. Eval Program Plann. 1982;5:233–7.64. Leape LL. Reporting of adverse events. N Engl J Med. 2002;347:1633–8.65. Avants SK, Margolin A, Holford TR, Kosten TR. A randomized controlledtrial of auricular acupuncture for cocaine dependence. Arch Intern Med.2000;160:2305–12.66. Hansten ML, Downey L, Rosengren DB, Donovan DM. Relationship betweenfollow-up rates and treatment outcomes in substance abuse research: moreis better but when is “enough” enough? Addiction. 2000;95:1403–16.67. Newell DJ. Intention-to-treat analysis: implications for quantitative andqualitative research. Int J Epidemiol. 1992;21:837–41.68. Little R, Yau L. Intent-to-treat analysis for longitudinal studies with drop-outs.Biometrics. 1996;52:1324–33.69. Yang Z, Hardin JW, Addy CL, Vuong QH. Testing approaches foroverdispersion in Poisson regression versus the generalized Poisson model.Biom J. 2007;49:565–84.70. Bellack AS, Bennett ME, Gearon JS, Brown CH, Yang Y. A randomized clinicaltrial of a new behavioral treatment for drug abuse in people with severeand persistent mental illness. Arch Gen Psychiatry. 2006;63:426–32.71. Comer SD, Sullivan MA, Yu E, Rothenberg JL, Kleber HD, Kampman K, DackisC, O'Brien CP. Injectable, sustained-release naltrexone for the treatment ofopioid dependence: a randomized, placebo-controlled trial. Arch GenPsychiatry. 2006;63:210–8.72. Fielding S, Fayers P, Ramsay CR. Analysing randomised controlled trialswith missing data: choice of approach affects conclusions. Contemp ClinTrials. 2012;33:461–9.73. Brooner RK, Kidorf MS, King VL, Stoller KB, Neufeld KJ, Kolodner K.Comparing adaptive stepped care and monetary-based voucherinterventions for opioid dependence. Drug Alcohol Depend.2007;88 Suppl 2:S14–23.74. Little RJ, D’Agostino R, Cohen ML, Dickersin K, Emerson SS, Farrar JT,Frangakis C, Hogan JW, Molenberghs G, Murphy SA, Neaton JD, Rotnitzky A,Scharfstein D, Shih WJ, Siegel JP, Stern H. The prevention and treatment ofmissing data in clinical trials. N Engl J Med. 2012;367:1355–60.75. Porta N, Bonet C, Cobo E. Discordance between reported intention-to-treatand per protocol analyses. J Clin Epidemiol. 2007;60:663–9.76. Wright CC, Sim J. Intention-to-treat approach to data from randomizedcontrolled trials: a sensitivity analysis. J Clin Epidemiol. 2003;56:833–42.77. Nosyk B, Guh DP, Bansback NJ, Oviedo-Joekes E, Brissette S, Marsh DC,Meikleham E, Schechter MT, Anis AH. Cost-effectiveness of diacetylmorphineversus methadone for chronic opioid dependence refractory to treatment.CMAJ. 2012;184:E317–28.78. Weinstein MC, Torrance G, McGuire A. QALYs: the basics. Value Health.2009;12:S5–9.79. Sculpher M, Fenwick E, Claxton K. Assessing quality in decision analytic cost-effectiveness models: a suggested framework and example of application.Pharmacoeconomics. 2000;17:461–77.80. Hoch JS, Briggs AH, Willan AR. Something old, something new, somethingborrowed, something blue: a framework for the marriage of healtheconometrics and cost-effectiveness analysis. Health Econ. 2002;11:415–30.81. Briggs A. Handling uncertainty in economic evaluation and presenting theresults. In: Drummond M, McGuire A, editors. Economic evaluation in healthcare: merging theory with practice. Oxford: Oxford University Press; 2001.p. 172–214.82. Nosyk B, Sharif B, Sun H, Cooper C, Anis AH. An economic evaluation andvalue of information analysis of three influenza vaccination strategies forpatients with HIV/AIDS. PLoS One. 2011;6, e27059.83. Brouwer WB, Koopmanschap M. How to calculate indirect costs ineconomic evaluations. Pharmacoeconomics. 1998;13:563–9.84. Swartz JA, Hsieh CM, Baumohl J. Disability payments, drug use andrepresentative payees: an analysis of the relationships. Addiction. 2003;98:965–75.85. Jepson RG, Harris FM, Platt S, Tannahill C. The effectiveness of interventionsto change six health behaviours: a review of reviews. BMC Public Health.2010;10:538.86. Klimas J, Field CA, Cullen W, O’Gorman CS, Glynn LG, Keenan E, Saunders J,Bury G, Dunne C. Psychosocial interventions to reduce alcohol consumptionin concurrent problem alcohol and illicit drug users: Cochrane review. SystRev. 2013;2:3.87. Hunt GE, Siegfried N, Morley K, Sitharthan T, Cleary M. Psychosocialinterventions for people with both severe mental illness and substancemisuse. Cochrane Database Syst Rev. 2013;10, CD001088.•  We accept pre-submission inquiries •  Our selector tool helps you to find the most relevant journal•  We provide round the clock customer support •  Convenient online submission•  Thorough peer review•  Inclusion in PubMed and all major indexing services •  Maximum visibility for your researchSubmit your manuscript atwww.biomedcentral.com/submitSubmit your next manuscript to BioMed Central and we will help you at every step:Richardson et al. BMC Public Health  (2016) 16:668 Page 16 of 16


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items