healthmedicinet

We have demonstrated that, with detailed planning and good communication with public
house staff, naturalistic studies of this nature can be run effectively. Communication
with individual proprietors of public houses was essential to keep them engaged with
the study. However, some important lessons were also learned. Communication with one
landlord broke down during the study, and he was not willing to participate further.
The dishwasher in one public house was too small to wash the experimenter’s curved
full-pint glasses, and the public house could not use them (see Table 1), so it used its standard glass range instead. The majority of customers during the
trial accepted the newly shaped glassware, although some did express dissatisfaction.
The collection of monetary takings from each landlord was straightforward and was
carried out by an experimenter after the first weekend and at the end of the study.
Overall, the brewer and landlords were satisfied with how the study was carried out
and there is potential for further collaboration.

The breakdown of communication with one landlord resulted in him not participating
on weekend 2 of the study. Experimenters were relying on the brewer to maintain good
communication with his landlord regarding the purpose and importance of the study.
Good communication is needed between the study team and each individual landlord directly
to foster loyalty to any future study. Experimenters need to engage each landlord
with the study and explain incentives to them effectively, so that potential benefits
to landlords are clear. Being a small public house with regular clientele, some of
them expressing dissatisfaction with the new glassware, may have been an important
factor in his decision not to participate on weekend 2.

The consensus from our post-study feedback with the brewer and landlords was that
no reimbursement for involvement was necessary except for being supplied with new
glasses at the end of the study. However, in our opinion, the brewer and landlords
would have been less keen to participate if there was not a second event that provided
publicity and increased custom for their public houses. If this type of incentive
is not feasible for a particular study, it may need to be replaced to encourage participation
from public houses in the future. This could include new glassware provided free of
charge, and/or the public houses’ participation could be publicised via blogs and
press releases after data collection has concluded. The optimal method of compensation
should be discussed with landlords. If any intervention serves to reduce monetary
takings, it would seem ethical and appropriate to offer some financial reimbursement
for this, given the competitive nature of the industry (particularly when participating
public houses are part of a relatively small local brewer, as in this case). Hopefully,
if a commitment to offer financial reimbursement for loss of earnings was made before
a future trial, this may alleviate the concern of landlords of maintaining equivalency
of earnings.

A public house’s dishwasher and shelving units should be inspected to ensure they
are fit for purpose before participation in any future study. Storing one public house’s
original stock of glasses off-site during the study was not an issue; however, if
more public houses are involved in future studies, this may be difficult. Where possible,
public houses should be able to store all glassware on-site. If this is not possible,
contingency plans should be put in place. The storage space of each public house should
be assessed at an early stage in future studies. If glasses are going to be given
to public houses as an incentive to participate in future studies, this may be less
of an issue. Assistance was provided by one experimenter to help stock and wash the
glasses which was appreciated by landlords. Further assistance from the study team
may be needed in future studies so that transfers of glassware are less disruptive.

There was a reduction in takings on weekends when straight glasses were stocked in
the three public houses compared to when curved glasses were stocked. Although the
mean difference (24 % reduction) was imprecise with wide confidence intervals, it
is worth noting that it was in the same direction as a previous laboratory study investigating
the effect of glass shape on the drinking rate of an alcoholic beverage 20]. The wide confidence interval (95 % CI 77 % reduction to 29 % increase) suggests
a large variability between the three public houses, and these results should be considered
with caution, as the study was not designed to support hypothesis testing. The main
purpose in collecting the results was to inform the design of a future full trial.
Dawkins Ales, which owns all three public houses, requested that monetary takings
not be made public, due to the commercial sensitivity of this information; therefore,
reporting aggregated results was not possible. However, if there was a larger number
of participating public houses involved, as would be the case in future studies, reporting
aggregate amounts of monetary takings would be possible with appropriate approval
because it would be harder to infer the takings of individual public houses in this
situation.

Results from the post-study investigation into the accuracy of monetary takings as
a proxy for alcohol consumption showed consistently strong, positive correlations
between units of alcohol sold and monetary takings for different beverage categories.
This suggests that monetary takings are an accurate proxy for alcohol consumption
based on these data. A limitation of the data used was the inability to calculate
the amount of alcohol sold on a daily basis within the 2-week window; therefore, a
correlation between units of alcohol consumed and total monetary takings could not
be calculated. A standard measure of alcohol units can be compared across sites to
evaluate alcohol use in different conditions (straight vs curved) provided that data
on units of alcohol in beverages of interest are available to researchers. In future
studies, landlords should be requested to send their takings (preferably broken down
by beverage type) on a weekly/periodic basis to experimenters via email or a collection
of them could also be arranged. Many modern tills have the capability to break down
purchases into different drink types, and this should be utilised in future studies.
This would allow sales of soft drinks to be accurately separated from alcoholic drinks
and changes in the sales of soft drinks to be monitored over the duration of a future
trial. Another option is to take an inventory of alcohol (e.g. number of kegs, bottles
etc.) although public houses may be reluctant to give such detailed information. If
this is the case, monetary takings can be used effectively as a proxy for alcohol
consumption. A balance must be kept between accuracy of alcohol consumption and maintaining
a naturalistic drinking environment.

Customers taking their custom elsewhere is a potential alternative explanation for
the reduction in alcohol consumption during the study when public houses were stocked
with straight glasses. This would appear to be a reduction in consumption in a trial,
but would not in fact reflect a reduction in individual-level consumption. This would
be difficult to monitor in a real-world environment. However, the low percentage of
patrons objecting to the straight glasses in this trial suggests that the level of
customer dissatisfaction may not deter public houses from participating especially
with the agreement to compensate for any lost profits during a trial.

Studies of this nature can be run on a relatively low research budget. Data collection
costs can be kept to a minimum at each study site, since the intervention can be delivered
within public houses and bars as part of their routine trade. Post-study feedback
suggested that two public houses found the intervention practical to implement. The
public house who withdrew from the study after weekend 1 found aspects of the study
impractical to implement, namely disruption caused by changing glassware and customer
dissatisfaction with experimenter-supplied glassware. Sensitivity to these types of
issues needs to be paramount when assessing the practicality of future study designs.

Glassware supplied by the study team did not include any branding, nucleation or volume
labelling, which are all common features on glassware used in on-trade premises. As
the intervention (unbranded) glassware was not used on two of the weekends, it is
possible that these aesthetic factors may have influenced alcohol consumption, rather
than the structural properties of the glass. This is a point to consider when trying
to integrate into the normal trade of a public house, since they may be reluctant
to stock glasses without these design features for a longer period, in which case
some elements (e.g. nucleation and volume labelling) may have to be applied to all
glasses. It should be considered that some consumers who expressed dissatisfaction
with experimenter-supplied glassware in this study may have done so due to the lack
of these common features of modern glassware being present. Considerations around
glass design (e.g. branding, nucleation) should be responsive to these views.

If the intervention is shown to be effective by reducing alcohol consumption, it would
need to be implemented legislatively, due to the demand to implement the intervention
voluntarily by public houses predicted to be low. In the UK, the 2003 Licensing Act
41] afforded powers to local licensing authorities to issue alcohol licences and enforce
the conditions of the licence in their area. This change has made licensing more local
and flexible to the needs of the local community. It has also made the process more
responsive to emerging evidence. Alcohol licensing conditions are not subject to the
same regulatory framework as, for example, treatments within the National Health Service,
meaning that evidence of efficacy can be directly translated into policy much more
rapidly. It is conceivable that the evidence from a future study could be implemented
in local authority licensing policies within 2–5 years of the end of the study (depending
on where in the licensing cycle the evidence becomes available). If results show that
straight glasses reduce consumption, a local licensing authority would be able to
add a requirement to stock straight glasses to its ‘menu’ of licensing conditions
which it can require premises to accept in order to be granted a licence. When an
existing premise applies to vary its licence, a responsible authority can demand certain
conditions to be met in order for the variation to be granted. Any person or responsible
authority (e.g. the local police force) can also apply to the licensing authority
for a review of an existing licence, with the aim of amending its conditions. If evidence
shows that straight glasses reduce consumption, the police or local licensing authority
may deem it worthwhile to require that more straight glasses be stocked in existing
licensed premises to bring about a reduction in crime and public disorder associated
with alcohol misuse 42]–44]. Critically, the intervention is one which, if mandated would not impose additional
direct costs on public houses and bars, since the glassware that constitutes the intervention
is no more expensive than the existing glassware. Moreover, since glassware is replaced
regularly (due to breakages etc.), any transition would have minimal impact.

A limitation of the study was that data on the usual business of the public houses
were not collected. However, we have no particular reason to think the weekends were
not representative of normal business when the study took place. Another limitation
of the study was that we explicitly targeted on-trade consumption of alcohol, but
individuals are increasingly consuming alcohol at home 1]. However, if consumption can be lowered in the on-trade market, this would still
have a significant impact on public health. Also, the hypothesised impact of straight
glasses is not exclusive to on-licence premises, and there is potential for a similar
effect in slowing drinking rate in the home.

Further studies should expand in scope to include other public houses over longer
periods of time to get a more comprehensive picture of the effect of glass shape on
alcohol intake. We suggest that the indirect measurement of alcohol consumption, using
monetary takings from itemised till receipts for alcoholic beverages, may be an appropriate
outcome measure in future studies. On the basis of our experience in this feasibility
study and the sample size calculation for a future trial, we estimate that a 6-month
data collection period in 30 public houses and bars would be sufficient to detect
a difference in monetary takings. Collaborating with larger chains of public houses
in the future would present unique challenges. The increased number of staff working
in these establishments would involve putting more trust in management to communicate
effectively with their employees. More glassware would be required, and a more substantial
logistical effort would be needed to transport and stock these glasses. Extra personnel
would be needed to carry this out. It may be more difficult to get larger chains of
public houses involved in public health research on their premises, given that stocking
straight glasses would impact on their business in the long term. It may be more fruitful
to engage with public houses with a community ethos rather than a high-volume business
model. Other key stakeholders, including local authorities and relevant trade associations,
have also pledged their support for future studies, and this should aid recruitment
efforts. Our research group is consistently forging links with industry partners which
will pay dividends when recruiting for future studies. To avoid attrition in future
studies, open communication should be maintained with each individual landlord so
that any issues and concerns can be dealt with as soon as they arise. A periodic meeting
between staff of public houses and experimenters during future studies is advisable.
Nevertheless, study designs should incorporate the possibility of attrition due to
participating public houses dropping out over the course of the study. Advantages
such as publicity and new glassware at low or no cost to their public house should
be emphasised to foster loyalty to future studies.