healthmedicinet

Included Studies

The initial search identified 3240 published studies. These were then divided equally
between four authors who screened the abstracts based on the inclusion criteria above.
The 164 papers identified in the initial screening were discussed by the authors with
the options of rejection, or possible inclusion into the review. At this point, 127
papers were rejected, with 37 for possible inclusion and further qualitative analysis.
After finding the full texts for the 37 papers, one further study was found by searching
through the citations of the included papers, and two very recently published studies
were sourced via correspondence with the authors, making a total of 40 papers for
initial qualitative synthesis (See Fig. 1).

Fig. 1. PRISMA Flow Diagram

The 40 included full text papers were then analysed and the following data extraction
tool was applied:

Type of study (RCT, observational etc.),

Tool used? (yes/no: if no then reject),

Primary/Secondary Care,

What is the tool?

What is the tool developed from? (for example the WHO checklist)

Is there evidence of the tool being used?

Has the tool been verified by others?

Has the tool led to any improvements?

After further analysis and meetings with the author team, only nine of the studies
were found to fit the inclusion criteria. The 31 excluded studies 17], 25]–54], along with justifications for exclusion, are presented in Additional file 3. The nine included studies were 51], 55]–62]. The characteristics of these studies are shown in Table 1. Please note that the original search was performed on 28.2.2014; however, six further
articles were subsequently identified via the peer review process and these were also
analysed and included in Fig. 1. The search for unpublished works (grey literature) did not yield any studies suitable
for inclusion in this review.

Table 1. Tools used in included studies

General Description of Included Studies

The nine included studies were found to have different outcome measures and therefore
different measures of success. These findings are summarised in Table 2.

Table 2. Outcome measures

Use of checklists

One recently published study 62] documented the introduction of a correct site surgery checklist to the dental division
of Central Manchester University Hospitals for outpatients undergoing dental extractions
under local anaesthesia, with or without sedation. The original checklist was introduced
to the division in January 2009; compliance with this checklist was poor, as revealed
by audit cycles. In 2012, a revised checklist was introduced based largely on the
WHO checklist. Clinical staff were engaged in the process of developing the policy
and training on the use of the checklist was provided to all members of staff and
students in the division. Following the introduction of the 2012 checklist policy,
audit revealed a 100 % compliance with the checklist. Importantly, no wrong tooth
extractions have occurred in the 24 months since the checklist was introduced. Prior
to this, during the period from 2009 to 2012, five cases of wrong tooth extraction
had occurred in the division.

Lee and colleagues 57], in an educational piece, advocated the development of an educational programme to
minimise the number of wrong tooth extractions occurring. The proposals were three
pronged; an educational programme, a pre-operative checklist and an informative unambiguous
referral form. It is important to note that this research was completed before the
creation of the WHO surgical safety checklist 63]. The authors suggest that one way of reducing error is to simplify the processes
involved. Their new referral form for practitioners is very straightforward, with
a schematic diagram of the mouth for the practitioner to mark with an ‘x’ the tooth/teeth
for removal. Following introduction of their guidelines for preventing wrong tooth/wrong
site surgery, the authors observed a reduction in the number of wrong tooth extractions
from 5 over a two-year period (before implementation of the policy) to nil over a
one-year period post-implementation.

A team based in Madrid 59] reported a proposal for an 18 point surgical safety checklist for use in patients
attending for ambulatory oral surgery procedures. The paper did not give details as
to whether adoption of the checklist has led to any decrease in the number of wrong
site procedures occurring in their department.

Another recently published study demonstrates the use of a safety checklist for patients
attending a specialist temporomandibular joint (TMJ) disorder clinic in Manchester
51]. In this study, the authors describe a checklist that is used to identify patients
with trismus, who are suffering from malignancy rather than TMJ related disorders,
whilst attending a clinic for TMJ disorders. The checklist is used when assessing
patients who present with trismus. The following diagnostic criteria will trigger
a review with a senior clinician and other investigations as appropriate:

Mouth opening of less than 15 mm,

Progressively worsening trismus,

Absence of a history of clicking,

Pain of non-myofascial origin,

Lymphadenopathy,

Presence of suspect intra-oral soft tissue lesions.

The authors describe two cases where misdiagnoses were made prior to the correct diagnoses
of malignancy. The checklist was introduced and one patient is subsequently described
as having been identified as having a malignancy via the use of this checklist. The
authors conclude that their checklist can be used to assist in the assessment of patients
with trismus, in order to avoid misdiagnosis of more sinister underlying pathology.

The four papers described are observational studies; three are before and after studies
and one is an educational piece; they do not make use of control groups. There are
clear statements of the findings from each study and all of the papers include a discussion
of the current policies and available literature. However, due to the lack of control
groups, their findings must be interpreted with caution.

Reporting Systems

We identified three papers that looked at the use of reporting systems in dentistry
58], 60], 61]. These papers have some overlap of authorship and are based around reporting systems
for adverse reactions to dental materials.

Lygre and colleagues 58] reported on a Norwegian National reporting system established in 1993. The aim was
to establish an adverse reaction registry and to serve as a clinical unit for patients
affected by reactions to dental materials. It was a voluntary reporting system, with
amalgam found to be the material most frequently causing adverse reactions (50 %).
It is not known whether this study led to any improvements in patient safety as it
is an epidemiological paper detailing the number of cases of adverse reaction, by
type over a six year period. This paper is frequently cited, although no subsequent
studies have validated their findings. The authors published a further paper two years
later 64], which demonstrated a reduction in symptoms in the patients who had restorations
replaced following adverse reactions.

The remaining two papers are from the UK. Scott and colleagues 60] sent reporting forms to dentists (27,000) and dental labs (2,700) encouraging responses.
A postal survey of 1,000 dentists also took place as authors suspected under reporting.
Amalgam was the most frequent cause of reaction for patients, with rubber products
and resins affecting dental staff and technicians respectively. Again, this was an
epidemiological paper with figures presented over a three year period of reporting.
Van Noort and colleagues 61] discuss the systems in place in the UK, Norway Sweden for reporting adverse reactions
to dental materials. Again, it is an epidemiological paper with details of 1,268 adverse
reaction reports from Norway, 848 from Sweden and 1,117 from the UK. Their findings
echo both the Scott and the Lygre studies in that the most frequent cause of adverse
reaction being metals (including amalgams) in patients, and rubber products in dental
professionals. They also found that there are no standardised criteria as to what
constitutes an adverse reaction to a dental material; and they also believed that
under-reporting was an issue. The authors acknowledged that it takes a considerable
amount of time for a pro-active reporting system to be established.

Through contact with the lead author (van Noort), it was established that the reporting
system is no longer in use due to a limited funding period, which has since expired.

The three papers analysed here are epidemiological studies with similar outcomes assessed
along with similar results. There has been sufficient analysis of the data found and
there are clear statements of findings and discussion around contemporary literature
and standards.

Use of Electronic Notes

Fricton 55] trialled an intervention that alerted the practitioner to web-based guidelines regarding
their patient’s medical conditions. In this three-arm, prospective, cluster randomised
clinical trial, two approaches were used to engage with practitioners. In the first,
a flashing alert was generated during the visit on the electronic notes system inviting
the practitioner to look at the current guidelines relating to the patient’s medical
conditions. In the second approach, patients who participated in the study were required
to ask their dentist to review the care guidelines specific to their medical conditions;
a control group was also used. Four medical conditions were included: xerostomia,
diabetes, chronic obstructive pulmonary disease (COPD) and congestive heart failure.
The authors found that the rate at which the practitioners accessed the guidelines
increased during the first six months, but that by the end of the study period (18 months),
the rate of use of guidelines had returned to the baseline levels. This may have been
due to the practitioners not feeling the need to continue reviewing the guidelines,
as they may have committed them to memory during the initial six month period, although
this hypothesis was not tested.

The trial did address a clearly focused issue, which was the use of a clinical decision
support tool integrated into electronic notes. Cluster randomisation was used to allocate
the clinics into the three arms of the trial; this is an appropriate method for this
type of trial and all providers were ‘blinded to’ the study protocol. The participating
practices were compared and found to be similar in terms of the number of people working
in them, the number of patients seen during the study period, and the number of patients
with the medical conditions being studied. Other than the intervention, the three
groups were treated the same. Several outcomes were measured with the key outcomes
being the number of website hits per dental care provider, the percentage of dental
care providers in each group accessing the online guidelines at the point of care
and the proportion of providers who continued to access the guidelines throughout
the study period. The authors found the statistically significant outcomes (P??0.05) to be an increase in the use of guidelines for all patients even if they
were not part of the study, and that provider activation was more effective than patient
evaluation. These results do suggest a precise estimate that the intervention was
successful.

The authors mention some significant limitations in the study; these were that 50 %
of the hits recorded were during appointments with patients who were not included
in the study (due to their funding source), and there may have been some crossover
between study arms as a number of providers worked at more than one clinic included
in the trial. It was also thought that there may have been crossover of patients between
clinics. These issues mean that the study was prone to bias in terms of allocation
and blinding of both patients and providers.

The effect of clinicians accessing these guidelines, or whether they have any impact
on patient safety outcomes, remains unclear.

Trigger Tools

A trigger is an easily detectable, focused item in a patient’s case notes that can
help to lead to the identification of an adverse event. Triggers can inform examination
of case notes to find out whether an adverse event actually occurred 65]. An example of a trigger is administration of naloxone, a drug that acts as an antagonist
to opioids. The use of naloxone implies that there has been an overdose of opioids
that would be considered an iatrogenic adverse event in the absence of drug abuse
or self-inflicted overdose.

One paper 56] discusses the use of trigger tools in dentistry. This paper is from the USA and the
authors compared the performance of an ‘Outpatient Adverse Event Trigger Tool’ modified
for dental clinic use with that of a review of randomly selected electronic patient
records. The study sites were dental practices where undergraduate dental students
provide treatment under supervision.

The dental trigger tools were described as follows:

1. Development of infections

2. Failure of complex procedures (for example implant failures)

3. Multiple visits – more than six completed visits during the six month review period,
or patients needing referrals to other specialists.

The electronic records were analysed for the above triggers via queries to the system.
Each record was also analysed by two dentists in order to determine whether any adverse
events had occurred by reading the narrative in the notes. The assessors also rated
severity. For further analysis, 50 records were randomly selected and analysed per
teaching dental practice to ascertain whether the findings from these notes were representative
of those from the larger sample.

In total, 8,931 patients were seen during the six month study period and the computer
system matched 315 cases with the triggers. Combining all triggers from the electronic
notes search compared to the assessor’s findings from the notes, they found that the
trigger tools had a positive predictive value of 0.5 (0.45-0.56, 95 % CI). The 50
randomly selected notes had a value of 0.34 (0.22-0.48, 95 % CI). In terms of severity,
the assessing dentists graded the adverse events observed. The majority were found
to have caused temporary harm, with one patient requiring hospitalisation due to infection
and nine having permanent harm (failed implants that were not replaced).

This paper had clear aims and results; the processes were described in detail and
the analysis was suitably rigorous. As the system for identifying the trigger tools
and adverse events was computerised, the authors attempted to compare the efficacy
of the system to their own analytic skills. We note that there was no assessment of
the inter-rater reliability of the two assessing dentists. The study concluded that
the dental clinic trigger tool was more effective in identifying adverse events than
was a review of randomly selected records.