Redesigning a large school-based clinical trial in response to changes in community practice.

2011·0.08 MB·English

by  Gerald, Lynn B

DESIGN Clinical Trials 2011; 8: 311–319
Redesigning a large school-based clinical trial in
response to changes in community practice
Lynn B Geralda, Joe K Geralda, Leslie A McClureb, Kathy Harringtonc, Sue Erwinc and
William C Baileyc
Background Asthma exacerbations are seasonal with the greatest risk in
elementary-age students occurring shortly after returning to school following
summerbreak.Recentresearchsuggeststhatthisseasonalityinchildrenisprimarily
related to viral respiratory tract infections. Regular hand washing is the most
effectivemethodtopreventthespreadofviralrespiratoryinfections;unfortunately,
achieving hand washing recommendations in schools is difficult. Therefore, we
designedastudytoevaluatetheeffectofhandsanitizeruseinelementaryschools
on exacerbations amongchildren with asthma.
Purpose Todescribetheprocessofredesigningthetrialinresponsetochangesin
thesafetyprofileofthehandsanitizeraswellaschangesinhandhygienepracticein
theschools.
Methods The original trial was a randomized, longitudinal, subject-blinded,
placebo-controlled, community-based crossover trial. The primary aim was to
evaluate the incremental effectiveness of hand sanitizer use in addition to usual
hand hygiene practices to decrease asthma exacerbations in elementary-age
children. Three events occurred that required major modifications to the original
study protocol: (1) safety concerns arose regarding the hand sanitizer’s active
ingredient; (2) no substitute placebo hand sanitizer was available; and (3)
communitypreferences changedregarding handhygienepracticesintheschools.
Results The revised protocol is a randomized, longitudinal, community-based
crossovertrial.Theprimaryaimistoevaluatetheincrementaleffectivenessofatwo-
step hand hygiene process (hand hygiene education plus institutionally provided
alcohol-based hand sanitizer) versus usual care to decrease asthma exacerbations.
Enrollment was completed in May 2009 with 527 students from 30 schools. The
intervention began in August 2009 and will continue through May 2011. Study
results should beavailable attheend of2011.
Limitations The changed design does not allow us to directly measure the
effectiveness of hand sanitizer use as a supplement to traditional hand washing
practices.
Conclusions The need to balance a rigorous study design with one that is
acceptable to the community requires investigators to be actively involved with
community collaborators and able to adapt study protocols to fit changing
community practices. Clinical Trials2011; 8: 311–319. http://ctj.sagepub.com
aMel and EnidZuckerman College ofPublicHealth,and theArizonaRespiratory Center, University ofArizona, Tucson,
AZ, USA, bDepartment of Biostatistics, University of Alabama at Birmingham, Birmingham, AL, USA, cLung Health
Center, University ofAlabama atBirmingham, Birmingham, AL, USA
Authorforcorrespondence:LynnBGerald,UniversityofArizona,MelandEnidZuckerman,CollegeofPublicHealth,
and theArizona Respiratory Center,Tucson, AZ, USA. E-mail: [email protected]
! The Author(s), 2011.
Reprints and permissions: http://www.sagepub.co.uk/journalsPermissions.nav 10.1177/1740774511403513
312 LB Gerald et al.
Background made to the original protocol due to both external
and local issues that arose after the original proto-
Poor asthma control in children is a well docu- col had been funded.
mented public health problem in the US [1,2]. It is
associated with frequent exacerbations that cause
respiratory symptoms, activity limitation, school Methods
absenteeism, and parental work absenteeism [3,4].
Exacerbations also lead to frequent urgent care Original study protocol
visits, emergency department visits, and hospitali-
zations [1,2]. Exacerbations are seasonal with the Study design
greatestriskamongelementary-agestudentsoccur-
ring shortly after returning to school following The original design was a randomized, longitudi-
summer break [5–8]. nal, subject-blinded, placebo-controlled, commu-
Outbreaks of viral respiratory infections have nity-based crossover trial. The primary aim was to
beenimplicatedasacauseforthisstrikingseasonal evaluate the incremental effectiveness of hand
pattern [8–13]. Regular hand washing is the most sanitizer use when added to a school’s typical
effective method to prevent the spread of viral hand washing practices to reduce asthma exacer-
respiratory infections [14,15]; unfortunately, bationsamongelementary-agestudents.Tocontrol
achieving effective hand washing practices in forschoolvariationandtominimizecross-contam-
schools is difficult [16,17]. Barriers include inade- ination, individual schools served as the unit of
quate time, insufficient soap or paper towels, and randomization. Crossover between intervention
inconveniently located sinks [18–20]. Assessment and placebo schools occurred during the summer
of schools’ physical environments has provided breakbetweenthefirstandsecondyear.Therewere
mixed results. A 1998 report of elementary school two major advantages of the crossover design. The
restrooms in the mid-Atlantic found that 66% of first was the ability to control for the strong
soap dispensers were nonfunctional or insuffi- seasonal variation in respiratory illnesses that was
ciently filled and 33% of automatic hand dryers expected. The second was the ability to maximize
were inoperable [21]. A 2009 study of primary and school support and participation because all
secondary school restrooms in New Mexico schools would eventually receive the intervention
reported that soap and hand drying were available instead of some being relegated to a noninterven-
in 90% of restrooms; however, hand sanitizer was tion control group.
reported in fewer than 2% [22].
To overcome perceived barriers associated with
hand washing, some schools have adopted antimi- Primary outcome
crobialrinse-freehandsanitizers[17,23–25].Recent
studies indicate that hand sanitizers reduce overall The primary outcome was the proportion of stu-
infection-related absenteeism among elementary dents who experienced an asthma exacerbation
school students by 20–50% [17,18,24–26], respira- each month defined as one or more of the follow-
tory illnesses by 30–50% [17,24,26], and teacher ing: (1) a red (<50% of personal best) or yellow
absenteeism by 10% [25]. Use of hand sanitizer in (50–70%ofpersonalbest)peakflowmeterreading,
the home has been shown to reduce asthma (2) increased use of quick relief medication from
exacerbations in children and respiratory illnesses baseline ((cid:2) 4 puffs), or (3) a respiratory-related
among family members [16,27]. Despite these school absence [29].
findings, a review by Meadows and Le Saux [28]
did not find sufficient evidence to specifically
recommend hand sanitizer use in general or Setting and recruitment
among children with asthma; however, a review
byJeffersonetal.[15]concludedthathandwashing A single large school district in Birmingham,
did reduce viral respiratory infections, particularly Alabama that was comprised of 30 elementary
in young children. schools with approximately 15,000 students was
To address the questions related to hand wash- recruited for participation. The system was racially
ing in general, and the benefit of hand sanitizer in (70%whiteand30%black)andeconomically(33%
particular, we designed a study to evaluate hand eligible for free or reduced lunches) diverse.
sanitizeruseinelementaryschoolsasamechanism Students with asthma were to be recruited by
to reduce exacerbations among children with school nurses who sent information packets home
asthma. We describe the study’s design, but more to parents via students. Interested parents were to
importantly, the process by which revisions were be encouraged to provide consent to be contacted
Clinical Trials 2011; 8: 311–319 http://ctj.sagepub.com
Redesigning large school-based clinical trials 313
by study staff or to contact the study staff directly. Gram-negative and Gram-positive bacteria associ-
Prior to obtaining written informed consent from ated with nosocomial infections and many viruses
parents and written assent from students, the associatedwithupperrespiratoryinfectionsinclud-
project was to be explained to both parents and inghumancoronavirusandadenovirus[32].Italso
students by the study staff. had been used previously in several large school-
basedstudieswithoutadverseevents[17,26,33,34].
Anotherimportantconsiderationwastheavailabil-
ity of a placebo without antimicrobial effects [17].
Study participants
Both alcohol-based and nonalcohol-based hand
Students who (1) attended one of the participating sanitizerswereinitiallyconsidered,butcommunity
schools, (2) had physician diagnosed asthma, and concerns regarding alcohol-based products and
(3)coulduseapeakflowmeterwereeligibleforthe their flammable nature [14,35] and the potential
study. for misuse made them a less desirable choice [36].
There were also preliminary data to suggest that
alcohol-based products might be inferior to non-
alcohol-based products [33,34]. For example, the
Intervention
antimicrobial activity of benzalkonium chloride
had been shown to increase over multiple, consec-
The intervention consisted of using a nonalcohol-
utive washes whereas the antimicrobial activity of
based hand sanitizer to supplement the schools’
ethyl alcohol tended to decrease over time [33,34].
typical hand washing practices. During the first
Both had better degerming activity than soap and
year, all schools were to receive hand sanitizer
water[33,34],but50%ofthoseusingethylalcohol
containing either an active ingredient or placebo
reported hand pain or discomfort whereas no one
that was dispensed from permanently mounted or
using benzalkonium chloride reported similar
free standing 1.8L dispensers (3600 applications)
symptoms [33].
thatprovidedamountsappropriateforelementary-
age students. All hand sanitizer was to be provided
by the study and was to be placed in the lunch-
roomsandclassrooms.Nosoap,papertowels,orair Data collection
dryers were to be provided to the schools.
Students, faculty, and staff in all schools would A web-based monitoring system (Asthma Agents)
be instructed on hand washing and hand sanitizer developed in collaboration with Blue Cross and
techniques based on the ‘Always Be Clean’ (ABC) Blue Shield of Alabama was to be used to collect
hand hygiene program developed by Woodward daily data without overburdening the schools or
LaboratoriesIncorporated.Handwashingwithsoap interrupting learning activities [29]. Peak flow
and water was to be promoted after using the meter (PFM) readings and school absences were to
restroomandwhenvisibledirtwasonthehandsas be recorded daily by students and verified by
recommendedbytheCDC[30].Handsanitizeruse teachersand/orschoolnurses.Quickreliefmedica-
was to be promoted as a supplement to hand tion (Proventil(cid:2) HFA) for in-school use was to be
washing upon arrival at the classroom, before providedatnochargetoallchildrenenrolledinthe
lunch, after using the restroom, at the end of the study by Merck and Company, Incorporated.
day, and after sneezing or coughing as recom- A DoserTM was to be attached to each student’s
mended by ABC. Grade-level appropriate instruc- inhaler to record each actuation of quick relief
tion was to be provided at the beginning of each medication automatically. The count was to be
school year with reinforcement on the first day of recorded every 2 weeks by the study staff.
each month. Faculty and staff were to be provided Compliance with hand hygiene recommenda-
continuing education prior to the start of each tionswastobeestimatedbasedonthefrequencyof
school year. refills of hand sanitizer in each classroom and in
the lunchroom. Refills were to be stored in the
housekeeping office at each school and the custo-
Hand sanitizer active ingredient dial staff recorded refill dates. This method was
chosen to minimize the burden on classroom
A nonalcohol-based hand sanitizer containing faculty and staff. Since all hand sanitizer was to
benzalkonium chloride (0.13%) was chosen for be supplied by the study, the total amount pro-
the study. Benzalkonium chloride, was designated vided to each school was to be monitored.
‘GRAS’ (generally regarded as safe) for topical Other explanatory variables and secondary out-
antiseptic applications by the Food and Drug comes including the student’s age, gender, race,
Administration (FDA) [31]. It was effective against asthmaseverity,qualityoflife,asthmacontrol,and
http://ctj.sagepub.com Clinical Trials 2011; 8: 311–319
314 LB Gerald et al.
household smoking exposure were to be collected of hand sanitizer use (i.e., flaky skin, lesions, rash,
during bi-annual phone interviews with parents. etc.) and how to report them.
Data analysis Major events and protocol revision overview
Data from the Asthma Agents system and the AfterreceiptofthegrantawardinAugust2007(See
DoserTMweretobeusedtocalculatetheproportion Table 1 for Study Timeline), three events occurred
of students in each group who experienced an that required major modifications to the original
exacerbation. Generalized estimating equations study protocol: (1) concerns arose regarding the
were to be used to model the marginal rate of safety of benzalkonium chloride; (2) no nonalco-
exacerbations, defined as the proportion of stu- hol-based hand sanitizer substitute was available;
dents within each school who experienced at least and (3) community standards changed regarding
one asthma exacerbation each month, while con- typical hand hygiene practices in the schools.
trolling for correlation between observations While responding to an IRB requirement to re-
within each student and between students within review product safety data following grant awards,
eachschool.Adjustmentforindividuallevelfactors we found new data indicating that benzalkonium
suchasthestudent’sage,race,gender,andasthma chloride:(1)inducedmoderategenotoxiceffectsin
severity were to be undertaken also. The study was eukaryotic cells [37], (2) produced histological
powered to detect a time averaged difference of changes (hyperplasia, incomplete keratinization,
7.5–10% between the exacerbation rates of the loss of the granular layer, acantholysis, and necro-
intervention and control schools given a sample sis)inorgan-culturedskin[38],(3)inducedbiofilm
size between 468 and 650 students with asthma. formation(amatrixofcellsattachingtoeachother
and a surface) of some bacterial species [39], and
(4) possibly caused allergic contact dermatitis [40].
Safety monitoring These concerns prompted a search for another
nonalcohol-based product but no substitute could
Thestudywasapprovedandtobemonitoredbythe be found. The only viable option was to switch to
Institutional Review Boards at the University of an alcohol-based product. While this alleviated
Alabama at Birmingham and the University of many of the safety concerns related to benzalk-
Arizona. A Data Safety and Monitoring Board onium chloride, it prevented the use of a subject-
(DSMB) was established to monitor adverse blindeddesignbecausetherewasnoplaceboforan
events. Two asthma safety events were mandated alcohol-based product.
as reportable: (1) a red PFM reading (<50% of Betweenthe time ofthe grant development and
personal best) with symptoms 3 days in a row and its award, the schools’ typical hand hygiene prac-
(2) use of more than 30 puffs of quick relief tices changed substantially due in large part to
medication in a 2 week period for those whose changes in the community’s perceptions regarding
medication was kept in the school office or more disease risk. School principals, faculty, and parents
than 40 puffs in a 2 week period for students who had become convinced that hand sanitizer was
self-carried. Self-carry of quick relief medication necessary for good hand hygiene within the
was allowed for any student who had the maturity schools. There was considerable fear that hand
to use it appropriately provided that the student’s washing alone was not sufficient to prevent ‘the
physician completed an authorization form. spreadofgerms’.Thisperceptionwasreinforcedby
Parents, teachers and school staff were to be amethicillin-resistantStaphylococcusaureus(MRSA)
providedhandouts describingpotentialsideeffects outbreak in several schools prior to the study.
Table1 Studytimeline
Date Schoolyear Activity
October2006 2006–2007 Submission
August2007 2007–2008 FundingAward
January–May2008 2007–2008 SchoolRecruitment
March2008–May2009 2007–2008 Parent/StudentRecruitment
2008–2009
August2009 2009–2010 InterventionStarts
August2010 2010–2011 Crossover
Clinical Trials 2011; 8: 311–319 http://ctj.sagepub.com
Redesigning large school-based clinical trials 315
Because of these concerns, hand sanitizer use only, we learned that personal hand sanitizer use
increased dramatically such that many schools hadbecomeubiquitousinmanyschools.Weasked
required parents to purchase personal hand saniti- schools to discourage personal hand sanitizer use
zer as a supply item. Despite reassurance by the (e.g., removing hand sanitizer from their required
local health department that hand washing supply lists) by arguing that hand washing was
remained the best option for hand hygiene, com- the preferred method of hand hygiene. The local
munity perceptions were so strong that a hand health department reinforced this message on
washing only arm was no longer a viable option. our behalf, but parents and school faculty coun-
tered that personal hand sanitizer use was sur-
mounting existing barriers to adequate hand
Revision process washing and that they were reticent to restrict its
use. They also pointed out the difficulty of enfor-
The decision to change a study protocol after a cingrestrictionsagainstpersonalhandsanitizeruse
grant award is a difficult one that involves many while at school.
important stakeholders. For this trial, important Unfortunately, we were unable to secure any
stakeholders included the National Heart, Lung, restrictions which meant that instead of the usual
andBloodInstitute(NHLBI)anditsdesignee(proj- care arm being a hand washing only arm, it would
ect officer), our Data Safety and Monitoring Board now also include a variable amount of personal
(DSMB), our Institutional Review Board, and the handsanitizeruseaswell.Theunexpectedincrease
community as represented by school administra- inpersonalhandsanitizerusepresentedapotential
tion,facultyandstaff,andparents.Ourgoalwasto biastothenulleffect;tocontrolforthispossibility
maintain the scientific integrity of the study while we developed instruments to monitor the schools’
adapting to safety concerns and changes in the hand hygiene practices more carefully.
community’s preferences. All of the revisions occurred prior to participant
Chronologically, the first major decision enrollment; therefore, no changes to the consent
involved the change from the original hand process were necessary.
sanitizer with its corresponding placebo to an
alcohol-based product without one. While the
safety concerns that prompted this decision were
mostly theoretical, we agreed that ensuring partic-
Revised study protocol
ipant safety, especially among children, warranted
erring on the side of caution. The principal inves- Revised study design
tigatorandkeystudypersonnelmadethisdecision
in close collaboration with the NHLBI project The revised protocol was a randomized, longitudi-
officer and DSMB over the course of multiple nal, community-based crossover trial that com-
teleconferences. The primary goal was to ensure pared the effectiveness of a standardized two-step
that the redesigned study addressed an important hand hygiene process (hand hygiene education
research question using a methodologically rigor- plus study-provided hand sanitizer and hand soap)
ous design without compromising student safety. with that of usual care (a variety of school-specific
The school system administration (superintendent hand hygiene practices).
and chief nursing officer) were appraised of the
rationale for the proposed changes and were asked
to approve them prior to implementation.
Because there was no placebo for the alcohol- Revised primary outcome
basedproduct,itwasnolongerpossibletoconduct
a subject-blinded design for a direct test of the No change.
effectiveness of hand sanitizer as a supplement to
hand washing. Instead, we chose to compare the
effectiveness of a standardized two-step hand
hygieneprocess(handwashingplushandsanitizer) Revised setting and recruitment
with handwashing only. Schools would be ran-
domly assigned to use study-provided hand saniti- No change.
zer and hand soap (intervention arm) or to hand
washing only (usual care arm). Crossover would
still occur after the first year.
Whenwemetwithschoolprincipalsandparent Revised study participants
advisory groups to discuss the revisions, including
randomization of some schools to hand washing No change.
http://ctj.sagepub.com Clinical Trials 2011; 8: 311–319
316 LB Gerald et al.
Revised intervention care schools did not receive any hand hygiene
educationoranyhandhygienesupplies.Theywere
The revised intervention consisted of a two-step expected to maintain their customary hand
hand hygiene practice that included regular hand hygiene practices. These practices varied across
washing with soap and water supplemented by schools and included hand washing and personal
hand sanitizeruse. Toovercome potentialresource hand sanitizer use of differing quality and fre-
barriers to adequate hand hygiene practices in quency. Some of this variation was attributable to
intervention schools, the study provided interven- the inconsistent supply of hand hygiene resources
tion schools with alcohol-based hand sanitizer, among some schools. Because of this variability,
handsoap,andneededrefills.Handsoapdispensers usual care schools might not be able to maintain
were installed by study personnel and hand the hand hygiene practices necessary to minimize
sanitizer was made available in disposable bottles thespreadofrespiratoryvirusesamongthestudent
in all intervention restrooms, health rooms, and population.
classrooms with a sink.
To our surprise the installation of dispensers Revised hand sanitizer/active ingredient
becameanimportantissue.Originally,weusedthe
installationofpermanenthandsanitizerdispensers Ethyl alcohol (62% solution) in foam (Purell(cid:2)
asanincentiveforschools toparticipate.Sincethe instant hand sanitizer foam).
hand sanitizer dispensers would hold product with
either an active ingredient or placebo, installing
them at the beginning of the study in all schools Revised data collection
did not pose a problem following crossover. After
the redesign, the lack of a placebo created the The data collection procedures for the primary
potential for a carry-over effect if Year 1 interven- outcome remained unchanged; however, two new
tion schools were to continue using the dispensers data collection instruments were designed to eval-
during Year 2 when assigned to usual care. We uate the school environment as it related to hand
proposed eliminating permanent dispensers alto- hygiene practices. The first instrument was used to
gether, but discontinuation became an issue with conduct a one-time baseline assessment of each
theschoolsastheywantedustohonorouroriginal school’s‘fixed’environmentalfacilitieswithregard
commitment. to the availability, type and working condition of
A compromise was reached by installing hand sinks, soap dispensers, hand sanitizer dispensers
soap dispensers, instead of hand sanitizer dis- and hand drying devices (i.e., paper towels or air-
pensers, in each school prior to the start of their dryers). The availability of hot and cold water and
intervention year. Schools were concerned that type of faucet mechanisms available (e.g., spring-
removingdispensersaftertheinterventionyearwas loaded,automatic,manual-turn)wasalsorecorded.
over might damage walls so we agreed to leave The second instrument was used for monthly
theminplacebutwewouldnolongersupplyrefills. monitoring of the availability of hand hygiene
Wedeemedthatthepotentialuseofthehandsoap resourcesateachschool.Thetypeofhandhygiene
dispensersduringtheusualcareyearposedlessofa supplies available at each school (usual care and
carry-over/contaminationriskthantheuseofhand intervention) and their availability, that is, liquid
sanitizer dispensers would. or bar soap, hand drying mechanism, hand saniti-
Inthereviseddesign,the‘AlwaysBeClean’hand zer, and hand hygiene instructions are recorded.
hygiene curriculum was replaced with the Centers General cleanliness and structural condition also
forDiseaseControlandPrevention’s(CDC)School were noted. Data were used to describe hand
Network for Absenteeism Prevention (SNAP) pro- hygiene resource availability in the usual care
gram (http://www.itsasnap.org/index.asp) because schools and to monitor implementation of the
the ABC program was a branded program of two-stephandhygieneprogramintheintervention
Woodward Laboratories Incorporated, the com- schools.
pany which produces the benzalkonium chloride Since hand sanitizer was recommended as a
hand sanitizer. Hand washing with soap and water supplement to standard hand washing in a setting
was promoted after using the restroom and when of limited school resources, it was deemed impor-
visible dirt was on the hands as recommended by tant to estimate the incremental value of hand
theCDC[30].Handsanitizerusewaspromotedasa sanitizer use. Therefore, an economic analysis was
supplement to hand washing upon arrival at the added as a study aim to estimate the additional
classroom, before lunch, after using the restroom, costs of hand sanitizer use. A careful cost account-
and after sneezing or coughing. ing of the usual care and intervention activities is
Only schools in the intervention arm received being conducted. Cost and effectiveness data will
the annual training and monthly refreshers; usual be combined to conduct a cost-effectiveness
Clinical Trials 2011; 8: 311–319 http://ctj.sagepub.com
Redesigning large school-based clinical trials 317
Table2 Revisedpowercalculations
Nocarryovereffect Carryovereffectpresent
(twoschoolyearscompared) (eachschoolyearcompared)
NumberofChildren Decreaseinfrequencyof Power(%) NumberofChildren Decreaseinfrequency Power(%)
PerSchool EPACs,dueto PerSchool ofEPACs,dueto
(30schools) intervention(%) (30schools) intervention(%)
14 7.5 74 14 7.5 48
10 92 10 71
15 7.5 77 15 7.5 50
10 93 10 71
16 7.5 79 16 7.5 54
10 94 10 74
EPAC¼EpisodesofPoorAsthmaControl.
analysis comparing the two-step hand hygiene Revised safety monitoring
intervention with usual care using dollars per
averted asthma exacerbation as the primary No change.
outcome.
Trial status
EnrollmentwascompletedinMay2009withatotal
Revised data analysis of527studentsenrolledin30schools(mean¼17.6
per school). The enrollment per school exceeded
Power calculations were revised based on the the estimate used to calculate the study’s power.
updatedstudydesignusingthemonthlyfrequency The intervention began in August 2009 and will
of episodes of poor asthma control (EPAC) as the end in May 2011.
primary outcome (Table 2). Two general scenarios
were assumed: one where there was no carryover
effects (data from both years) and one where there
Conclusions
were such effects (each year evaluated separately).
GEEwasusedtoaccountforthecorrelationwithin
One of the most difficult aspects of the academic-
a child over time as well as the correlation among
community partnership is the lag time between
children within the same school.
study development, funding, and implementation.
Carryover (or residual) effects occur when the
During this interval, key stakeholders, perceptions,
effect of an intervention during the first time
and standard practices within the community can
period persists into the second time period.
change,ofteninanunpredictablemannerwhichis
Thepersistenceoftheeffectmaybiasthemeasured
outside the control of the research team. At the
effectofthesecondintervention.Itispossiblethat
same time, new scientific discoveries may emerge
educational intervention could lead to sustained
that render the original study assumptions obso-
changesinhandhygienepracticesatinterventions lete. Because of this dynamic environment, the
schools that might persist into the second year; community-basedresearchermustberesponsiveto
however, the presence of a summer break provides changes that may undermine the original study
a 10-week ‘wash out’ period that should minimize design orcollaborations withcommunity partners.
any carryover effect since no reinforcement is We have found that a study team with a broad
provided to original intervention schools during range of skills, talents, and interests is well posi-
the second year. Research has shown that changes tioned to respond to a quickly changing environ-
inbehavioralpracticestypicallydeclineby40–80% ment. Not all team members are equally suited to
after 6 weeks without reinforcement [41]. A recent serve as liaisons with community partners. In
reviewofhandhygieneinterventionsinhealthcare school-based research, team members with prior
workers noted that without reinforcement the workexperienceintheschools(e.g.,formeradmin-
changes in hand hygiene behaviors were sustained istrators, teachers, or coaches) often can establish
forlessthan1week[42].A1monthbaselineperiod more effective collaborations with school person-
atthebeginningofeachschoolyearwasplannedto nel than ‘academicians’. These team members can
measure potential carryover effects. usetheirpreviousrelationshipstoidentifyandgain
http://ctj.sagepub.com Clinical Trials 2011; 8: 311–319
318 LB Gerald et al.
access to the key stakeholders within the school compromise the ethical conduct of research, parti-
system. Additionally, they can navigate a cultural cularly in vulnerable populations. It also may be
workenvironmentthatisoftenquitedifferentfrom counterproductive as the community partners are
the academic setting. Having team members with vital to the translation of research knowledge into
this background is critically important for school- sustainable practices.
based research.
In retrospect, the most crucial decision that
impacted our original research design was the
Funding
decision to abandon the nonalcohol-based hand
sanitizer containing benzalkonium chloride.
This work was supported by a grant from the
Without this product and its corresponding pla-
National Heart, Lung, and Blood Institute 1 R01
cebo, we were unable to conduct a subject-blinded
HL086972. Quick relief medication for in-school
design.Initially,wefeltconfidentthattheproduct
use was donated by Merck and Company, Inc. and
was safe and effective; however, the in vitro data
datasystemsupportwasprovidedbyBlueCrossand
that subsequently emerged cast doubts on the
BlueShield of Alabama. Registered at clinicaltrials.-
product’s previously established safety record. The
gov NCT00528814.
new data emerged at a point in the study cycle
where little time was available to contemplate
the decision without substantially delaying imple-
mentation. We also were sensitive to safety con- References
cernsgiventhatwewereworkingwithavulnerable
population (children) in a challenging research 1. Akinbami LJ, Schoendorf KC. Trends in childhood
setting (schools). We believed that a smaller body asthma:prevalence,healthcareutilization,andmortal-
ity.Pediatrics2002;110:315–22.
ofin vitro findingsquestioning theproduct’ssafety
2. Akinbami L. The state of childhood asthma, United
trumped a larger body of in vivo findings demon-
States,1980-2005.AdvData2006;381:1–24.
stratingsafetythatwereobtainedthroughindustry- 3. Schmier JK, Manjunath R, Halpern MT, et al. The
funded, manufacturer-led research. impact of inadequately controlled asthma in urban
The impact of eliminating the subject-blinded childrenonqualityoflifeandproductivity.AnnAllergy
AsthmaImmunol2007;98:245–51.
design was magnified by the abrupt change in
4. Bateman ED, Frith LF, Braunstein GL. Achieving
community hand hygiene practices within the guideline-based asthma control: does the patient bene-
schools. If personal hand sanitizer use had not fit?EurRespirJ2002;20:588–95.
been prevalent in usual care schools, the crossover 5. Dales RE, Schweitzer I, Toogood JH, et al. Respiratory
infectionsandtheautumnincreaseinasthmamorbidity.
design would have been a robust test of hand
EurRespirJ1996;9:72–77.
sanitizer as compared with hand washing alone.
6. Silverman RA, Ito K, Stevenson L, Hastings HM. The
However, the ubiquitous nature of personal hand relationship of fall school opening and emergency
sanitizer use forced us to rely on what had been department asthma visits in a large metropolitan area.
reported previously to be substandard hand ArchPediatrAdolescMed2005;159:818–23.
7. Blaisdell CJ, Weiss SR, Kimes DS, et al. Using seasonal
hygiene practices within the schools due to signif-
variations in asthma hospitalizations in children to
icantstructuralbarriers.Bysupplyingtheeducation predict hospitalization frequency. J Asthma 2002; 39:
andresourcestoensurethatthebarrierstoeffective 567–75.
hand hygiene in the intervention schools were 8. Johnston NW, Johnston SL, Ducan JM, et al. The
September epidemic of asthma exacerbations in chil-
eliminated, we were able to evaluate the best
dren: a search for etiology. J Allergy Clin Immunol 2005;
possible hand hygiene practice (hand washing
115:132–38.
plus hand sanitizer use) against a variety of usual 9. Freymuth F, Vabret A, Brouard J, et al. Detection of
care practices. Unfortunately, the study no longer viral, Chlamydia pneumoniae and Mycoplasma pneu-
canmeasure theeffectiveness ofhandsanitizeruse moniae infections in exacerbations of asthma in chil-
dren.JClinVirol1999;13:131–39.
directly as a supplement to traditional hand wash-
10. Johnston SL, Pattemore PK, Sanderson G, et al.
ing practices. If the trial results are negative, the Community study of role of viral infections in exacer-
possibility of bias created by personal hand saniti- bations of asthma in 9-11 year old children. BMJ 1995;
zer within usual care schools may be the primary 310:1225–29.
11. MontoAS.Theseasonalityofrhinovirusinfectionsand
contributing factor.
its implications for clinical recognition. Clin Ther 2002;
Community-based clinical trials are needed to
24:1987–97.
establish the effectiveness of interventions under 12. Rakes GP, Arruda E, Ingram JM, et al. Rhinovirus and
real-world conditions [43]. A methodologically respiratory syncytial virus in wheezing children requir-
rigorous study design is important to achieve this ing emergency care. IgE and eosinophil analyses. Am J
RespirCritCareMed1999;159:785–90.
goal, but the community-based researcher must
13. Thumerelle C,Deschildre A,Bouquillon C, etal. Role
also respect community beliefs and practices [44].
of viruses and atypical bacteria in exacerbations of
Failure to address them satisfactorily may asthma in hospitalized children: a prospective study in
Clinical Trials 2011; 8: 311–319 http://ctj.sagepub.com
Redesigning large school-based clinical trials 319
theNord-PasdeCalaisregion(France).PediatrPulmonol 29. Gerald LB, McClure LA, Mangan JM, et al. Increasing
2003;35:75–82. adherence to inhaled steroid therapy among school-
14. CentersforDiseaseControlandPrevention.Guideline children: randomized, controlled trial of school-based
for hand hygiene in health-care settings: recommenda- supervisedasthmatherapy.Pediatrics2009;123:466–74.
tions of the healthcare infection control practices advi- 30. Centers for Disease Control and Prevention. Clean
sorycommitteeandtheHICPAC/SHEA/APIC/IDSAHand Hands Saves Lives!; 2010, http://www.cdc.gov/hand
HygieneTaskForce.MMWR2002;51:1–44. washing/(accessed22March2011).
15. Jefferson T, Del Mar C, Dooley L, et al. Physical 31. DepartmentofHealthandHumanServicesFoodand
interventions to interrupt or reduce the spread of Drug Administration. Tentative final monograph for
respiratory viruses: systematic review. BMJ 2009; 339: health-care antisepticdrugproducts: proposed rule.Fed
b3675. Regist1994;59:31441–52.
16. Lee GM, Salomon JA, Friedman JF, et al. Illness 32. Wood A, Payne D. The action of three antiseptics/
transmission in the home: a possible role for alcohol- disinfectants against enveloped and non-enveloped
basedhandgels.Pediatrics2005;115:852–60. viruses.JHospInfect1998;38:283–95.
17. WhiteCG,ShinderFS,ShinderAL,DyerDL.Reduction 33. DyerDL,GerenraichKB,WadhamsPS.Testinganew
of illness absenteeism in elementary schools using an alcohol-freehandsanitizertocombatinfection.AORNJ
alcohol-free instant hand sanitizer. J Sch Nurs 2001; 17: 1998;68:239–241,243–234,247–251.
258–65. 34. Moadab A, Rupley KF, Wadhams P. Effectiveness of a
18. Guinan M, McGuckin M, Ali Y. The effect of a nonrinse,alcohol-freeantiseptichandwash.JAmPodiatr
comprehensive handwashing program on absenteeism MedAssoc2001;91:288–93.
in elementary schools. Am J Infect Control 2002; 30:
35. BryantKA,PearceJ,StoverB.Flashfireassociatedwith
217–20.
the use of alcohol-based antiseptic agent. Am J Infect
19. Robinson J. The changing role of the school nurse: a
Control2002;30:256–57.
partnerininfectioncontrolanddiseaseprevention.JSch
36. RobertsHS,SelfRJ,CoxonM.Anunusualcomplication
Nurs2002;Suppl.:12–14.
ofhandhygiene.Anaesthesia2005;60:100–101.
20. VesseyJA,SherwoodJJ,WarnerD,ClarkD.Comparing
37. FerkF,MisikM,HoelzlC,etal.Benzalkoniumchloride
handwashingtohandsanitizersinreducingelementary
(BAC) and dimethyldioctadecyl-ammonium bromide
school students’ absenteeism. Pediatr Nurs 2007; 33:
(DDAB), two common quaternary ammonium com-
368–72.
pounds, cause genotoxic effects in mammalian and
21. Early E, Battle K, Cantwall E, et al. Effect of several
plant cells at environmentally relevant concentrations.
interventions onthefrequencyof handwashingamong
Mutagenesis2007;22:363–70.
elementary public school children. Am J Infect Control
38. VaraniJ,PeroneP,SpahlingerDM,etal.Humanskinin
1998;26:263–69.
organ culture and human skin cells (keratinocytes and
22. Ramos MM, Blea M, Trujillo R, Greenberg C.
fibroblasts) in monolayer culture for assessment of
Inspectionsofhandwashingsuppliesandhandsanitizer
chemically induced skin damage. Toxicol Pathol 2007;
inpublicschools.JSchNurs2010;26:393–97.
23. WhiteC,KolbleR,CarlsonR,etal.Theeffectofhand 35:693–701.
39. Houari A, Di Martino P. Effect of chlorhexidine and
hygiene on illness rate among students in university
residencehalls.AmJInfectControl2003;31:364–70. benzalkonium chloride on bacterial biofilm formation.
24. Morton JL, Schulta AA. Healthy hands: use of alcohol LettApplMicrobiol2007;45:652–56.
gelasanadjuncttohandwashinginelementaryschool 40. Hann S, Hughes TM, Stone NM. Flexural allergic
children.JSchNurs2004;20:161–67. contactdermatitistobenzalkoniumchlorideinantisep-
25. Hammond B, Ali Y, Fendler E, et al. Effect of hand ticbathoil.BrJDermatol2007;157:795–98.
sanitizer use on elementary school absenteeism. Am J 41. MarlattGA,GordonJR.RelapsePrevention:Maintenance
InfectControl2000;28:340–46. StrategiesintheTreatmentofAddictiveBehaviors.Guilford
26. Dyer DL, Shinder A, Shinder F. Alcohol-free instant Press,NewYork,1985.
hand sanitizer reduces elementary school illness absen- 42. Naikoba S, Hayward A. The effectiveness of interven-
teeism.FamMed2000;32:633–38. tions aimed at increasing handwashing in healthcare
27. SandoraTJ,TaverasEM,ShihMC,etal.Arandomized, workers – a systematic review. J Hosp Infect 2001; 47:
controlledtrialofamultifacetedinterventionincluding 173–80.
alcohol-based hand sanitizer and hand-hygiene educa- 43. Flay BR. Efficacy and effectiveness trials (and other
tiontoreduceillnesstransmissioninthehome.Pediatrics phases of research) in the development of health
2005;116:587–94. promotionprograms.PrevMed.1986;14:451–74.
28. Meadows E, LeSaux N. A systematic review of the 44. IsraelBA,SchultzAJ,ParkerEA,BeckerAB.Reviewof
effectiveness of antimicrobial rinse-free hand sanitizers community-based research: assessing partnership
forpreventionofillness-related absenteeisminelemen- approaches to improve public health. Ann Rev Pub
taryschoolchildren.BMCPublicHealth2004;4:50. Health1998;19:173–202.
http://ctj.sagepub.com Clinical Trials 2011; 8: 311–319