diagnosis - monitoring - diabetic neuropathies
diagnóstico - monitoramento - neuropatias diabéticas
Diabetic sensorimotor polyneuropathy (DSPN) affects approximately 10% of the subjects
newly diagnosed[1],[2],[3] as type 2 diabetes and this percentage can increase by duration, lack of glycemic
and cardiovascular complications[3],[4],[5]. When appropriately treated by restoration of glycemic control[3], the progression of the DSPN can be delayed and the diabetic ulcers and amputations
reduced[3],[6]. However, the neuropathy is still the major cause of diabetic foot presenting damaged
nerve fibers and this secondary complication affects more than 50% among the subjects
diagnosed from long time, despite could be appreciated that DSPN does not surspass
the 10% hallmark[7].
Different guidelines recommend annual evalutation for DSPN and clinical examination
of the lower extremities and feet in subjects with diabetes, representing a significant
burden for basic care services whereas primary prevention[1],[2],[8].
Regularly, the gold standard methods to diagnose DSPN, e.g. nerve condution studies,
do not are always easily avalible due the high cost, especially in Brazil[7]. Given that, the Michigan Neuropathy Screening Instrument (MNSI) has been described
as an alternative low cost method by fast application which allows to score, classify
and diagnose the neuropathy[9].
Developed at Michigan Diabetes Research and Training Center in the United States,
the MNSI aims to screen the symmetric diabetic neuropathy from individuals with diabetes
mellitus and its reliability and acuracy were discussed in previous studies[9],[10],[11], but not for Brazilian population.
Also available to screen peripheral neuropathy, the Pain Quality Assessment Scale
(PQAS) were cross-adapted for Brazilian Portuguese; however the PQAS was designed
to focus on assessing the quality of the neuropathic pain in cancer patients[12]. Due to the specific approach to DSPN that includes other issues related to peripheral
neuropathy apart from neuropathic pain, the MNSI could be considered a better scale
to be used in the diabetic population.
Given the circumstances and once authorized by the MNSI creators, this study aimed
to adapt cross-culturally the MNSI into Brazilian Portuguese, verifing its reliability.
METHOD
In order to ensure the quality of the adapted tool, we carried out a cross-sectional
study to translate and adapt transculturally the MNSI, called by us MNSI-Brazil. The
essential steps to accomplish our aim was guided by the process to adapt transculturally
self-report measures published by Beaton and collaborators[13] in five sequential stages: 1) translation, 2) synthesis, 3) back translation, 4)
expert committee review and 5) pretesting, accompanied paralelly by the transversal
stage of submission and appraisal of all written reports by developers.
This guideline refers to international rules established to secure the equivalence
maintenance between the original questionnaire version and the destiny, in this case:
the Brazilian population. Once finished, the MNSI-Brazil was submitted to psychometric
testing to verify the inter-rater and inter-test reliabilities.
This study was approved by Fundação de Ensino e Pesquisa em Ciências da Saúde (FEPECS)
Ethics Committee, D.C. (Report 160.752/2012) and all the participants signed a term
of free and enligthened acceptance and they were informed about the procedures of
all the stages in the study.
The Michigan Neuropathy Screening Instrument
The original version of MNSI was created in the Michigan Center for Diabetes Translational
Research (MCDTR) and the authorization to performe the cross-cultural adaptation was
given by Pamela A. Campell as requested by e-mail sent May 21, 2013 and answered May
22, 2013.
The instrument is composed by an introduction which gives guidance on how to use MNSI,
followed by two application forms where one of them is self-administered by the patient.
The first one, the self-administred, has been prepared to score the clinical history
(history questionnaire) and the second one to score physical assessment. Afterward,
the individual scores are added up to give a total value. Total value larger than
8 suggests a symmetrical peripheral neuropathy.
Translation
All parts of the original MNSI (including the introduction) were translated into Brazilian
Portuguese by two native Brazilian Portuguese speakers who worked independently. They
were fluent in English language and have different profiles and academic education
areas (physical therapy and engeneering). One of them was the “ingenuous” translator,
because he did not have experience in the health science. The used “ingenuous” strategy
obtains a translation which reflects the linguistic norms practiced by population
without be influenced by schoolar formality.
Synthesis
In order to prepare the first Brazilian Portuguese version, the authors compared and
synthesized the two translations by consensus. The two translations were consistent
in almost its totality. Only three text fragments (words or phrases) chosen by each
translator were different, however expressing the same meaning from the Brazilian
Portuguese lexicon. At the end, they were defined by consensual decision made by the
authors in: 1) “classificação” instead of “t
riage”, 2) “perguntas sob
re a sensibilidade d
e suas pernas e pés instead of questões
sobre a sensiblidade de suas
pernas e pés” e 3) “hálux
” instead of “dedo grande do pé”.
Back translation
The first Brazilian Portuguese version was back translated into English by two professional
bilingual translators who were fluent in Portuguese and English. They did not participate
in the previous stage and did not know the MNSI.
Following, the two back translated versions were compared with the original MNSI to
validate the consistence in the translated version which reflected the same original
meaning. The two back translations were very similar with just two differents text
fragments. As in the synthesis, by consensual decision were defined: 1) “Michigan
Neuropathy Screening Instrument” instead of “Michigan Neuropathy Classification Instrument”
and 2) “numb” instead of “asleep”.
Expert committee review
A committee composed by 3 rehabilitation and health specialists, bilingual, was assisted
by the first author to achieve cross-cultural equivalence and consolidate all the
versions of the questionnaire and develop what was considered the prefinal version
for field testing.
Committee’s meetings were regularly performed to seek the linguistic equivalence necessary
to make the prefinal version[13]. The words julged do not be equivalent by one of the members and text adaptations
were reviewed and discussed to reach agreement on the preliminary version applied
to Brazilian population.
Pretesting
The prefinal version ([Appendix 1, 2 and 3]) of the MNSI-Brazil was applied in 30 subjects with diabetes mellitus and tested
to assure cross-cultural equivalence. During the pretesting, we did not find words
of which Brazilian attendee could not understand. So the MNSI-Brazil final version
did not change from de prefinal version.
The 30 subjects presenting clinical diagnosis of type 2 diabetes mellitus were evaluated
by MNSI-Brazil. The sample was formed by convenience from two health services: 1)
an assistential program called Universidad
e Aberta à Terceira
Idade (UNATI) linked to the Universidade
Estadual de Goiás (UEG) located in Goiânia, Goiás, Brazil and 2) a Center Health Service (CHS) called
CHS number 3, located in the Ceilândia Administrative Area, Distrito Federal, Brazil.
Participants presenting peripheral or central neurological-related illness do not
associated with diabetes in the medical records (e.g. traumatic injuries, infections,
inherited causes and exposure to toxins) and those ones with cognitive problems identified
by Mini-Mental State Examination (MMSE) were excluded. Depending the scholar level,
we consider different scores as MMSE cut-off point as recommended. The 13, 18 and
26 scores respectively for those not able to read and write, for those attending at
least 7 years of scholling and those schooling 8 years or more[14],[15]. In order to give general antropometric characteristics of the sample, we included
information about body composition (body mass index) and skin color (leucoderm, faioderm
and melanoderm).
The MNSI-Brazil was applied in an acclimatized room where the volunteeers were confortably
accommodated to aswer the questions put by the raters who filled out the instrument.
They spended the time required and the raters were oriented to take notes from dubious
words or unclear answer. The time performed to apply the MNSI-Brazil was timed by
each rater and all participators. At the end, the participants were questioned about
their difficulties. The pretesting was conducted from August 2014 to May 2015.
Inter-rater and inter-test reliability
A group composed by 22 volunteers among the initial sample (n = 30) agreed to take
part in the reliability tests. The tests were performed in three times (trial 0, 1
and 2), during four different days (day 1, days 2 or 3 and day 20), by two different
raters: rater 1 and 2. The 30 initial volunteers completed the first MNSI-Brazil application
oriented by the rater 1. In the day 2 or 3, 22 among them repeat the test applied
by other rater to test the inter-rater reliability. Twenty days after the test, the
rater 1 retested the 22 volunteers.
The inter-rater reliability was verified during the test phase comparing the MNSI-Brazil
scores took in the trial 0 (day 1) and trial 1 (days 2 or 3). The trials 0 and 1 was
performed by different raters (they performed blind-independent assessments). In turn
to verify inter-test reliability, the scores obtained in the trial 0 (day 1) were
compared to those same scores took during trial 2 (day 20) by the rater 1. After 20
days is few probable to find peripheral nerve degeneration with visible clinical evaluation
and it is a good period to confirme inter-test repeatability[16].
Data processing and statistical analysis
The Shapiro Wilk normality test was used to verify whether the variables demonstrated
the Gaussian distribution and it determined the need for parametric tests in the analysis.
The significance level for all analyses was established at alfa equal 0.05. We use
descriptive statistic to characterize the sample used to reliability tests by mean
of average, standart deviation (SD) and frequency distribution (absolut and relative).
The Intra-class Correlation Coefficient (ICC) and the Limits of Agreement (LOA) were
used to define the respective quality and magnitude of inter-rater and test-retest
reliability, which were plotted with a 95% confidence interval (CI) using the Bland
Altman method. ICC values above 0.75, between 0.40 and 0.75 and below 0.40 represented
excellent, moderate and poor reliabilities respectively[17].
RESULTS
Translation and cross-cultural adaptation
We identify few semantic, linguistic or cultural diferences during the process of
MNSI translation and none serious discrepancies about the vocabulary. Those few divergent
words and adaptation in the text were elucidated and eliminated by the expert committee
review. The same occurred in the back translation stage. During the pretesting, all
questions were appropriately answered and comprehended for the totality of the participants.
During the translation stage, between the versions made by the two translators, we
do not found divergencies, but rather synonym words. Then, this point was solved choosing
the more common synonymous used by Brazilian people.
In the same way, the back translation presented few divergencies between the two versions
made by each English language native translators and the choice was done by the terms
used in the original version.
Reliability tests
The measurement conditions were defined by the raters who tested (rater 1 and 2) the
volunteers and by the moment when the measurements were obtained (test and retest),
establishing two mensurement conditions to test inter-rater and inter-test reliabitily.
The participants were 69.05 ± 7.59 years old (mean ± standard deviation), within the
overweight range defined by adult Body Mass Index (BMI) classification (28.56 ± 4.05).
All participants were presenting adequated cognitive status scored evaluated by MMSE
in 26.55 ± 3.17 and verified during the tests by the full comprehension of the self-administered
part of the MNSI-Brazil. All participants were able to read and write (36% attending
at least 7 years of scholling and 64% schooling 8 years or more).
The sample retested was predominantly female (95%) and composed by leucoderm, faioderm
and melanoderm people as established by Edgard Roquette Pinto who proposed a system
to classify the three main categories based in the skin color and present in the Brazilian
population ([Table])[12]. The time spent to apply the self-administered part of the MNSI-Brazil was around
three minutes.
Table
Characteristics of the group submitted to reliability tests.
|
Variables
|
Average or n
|
SD or %
|
|
Quantitative
|
|
Age (years old)
|
69.05
|
±7.59
|
|
BMI
|
28.56
|
±4.05
|
|
Mini-mental (score)
|
26.55
|
±3.17
|
|
MNSI-Brazil (rater 1)
|
4.36
|
±2.15
|
|
MNSI-Brazil (rater 2)
|
4.22
|
±2.24
|
|
MNSI-Brazil (retest)
|
4.04
|
±1.96
|
|
Qualitative
|
|
Men
|
1
|
5%
|
|
Women
|
21
|
95%
|
|
Faioderm
|
11
|
50%
|
|
Leucoderm
|
6
|
27%
|
|
Melanoderm
|
5
|
23%
|
|
Group from the total sample (n = 30)
|
22
|
100%
|
Quantitative variables are presented in mean ± standard deviation (SD) and qualitative
variables are in absolut (n) and relative (%) frequency distribution. BMI: Body Mass
Index; MNSI: Michigan Neuropathy Screening Instrument.
The average measured by the MNSI-Brazil in the 3 measurement conditions (rater 1,
rater 2 and retest) was 4.21 ± 2.09, ranging from 0 to 8 total scores. Just two subjects
were scored 8 or more defining a diagnosis suggestion of DSPN. In the first measurement
condition (rater 1) one subject had been scored 8 (DSPN diagnosis suggestion), however
the dignosis was not suggested by the rater 2 and in the retest. Otherwise, the other
subject who was scored 7 by the rater 1 was scored 8 by the rater 2 and retested confirming
the diagnosis suggestion of DSPN.
The analysis of the inter-rater and inter-test reliability ([Figure]) obtained by the difference (Bland Altman method) and ICC values calculated from
measurements taken by each rater (ICC = 0.840) and in each test (0.864) respectively
indicated excellent repeatability[17]. The Bland Altman plots detected mean differences between raters and tests showing
small or no significant deviations from zero for most of the subjects. In general,
the 95% LOA that was obtained between the different measures ranged from -2 to 2 points
(upper and lower graphs in the [Figure]).
Figure The bland-altman plots.
DISCUSSION
To date, none cross-cultural adaptation had been done to the MNSI. Then, aside from
the newness to introduce a Portuguese version of this clinical tool, this study tested
the reliability of the measuments obtained from this diagnosis method in the Brazilian
population with diabetes mellitus. The cross-cultural adaptation avoids multinational
and multicultural diversity to uniformize the concepts and evaluation aspects desired
with losing the principles of the diagnosis method.
In order to make available the instruments created around the world, the cross-cultural
adaptation of the measurement instruments related to human health is essential to
use it in other countries beyond the country where it was developed in a safe and
equivalent way to original questionnaire[13].
Nowadays, the DSPN is diagnosed by mean of clinic signals and information obtained
from many complementary exams demanding time and procedures to conclude the diagnosis.
Then, a unique instrument fast and easy-to-apply would be helpful in the everyday
clinical work. The MNSI-Brazil can suggest DSPN diagnosis by a set of clear points
in the questionnaire and simple physical examination to compose a score system by
which we can define objectively if the patient may have or not DSPN.
The Portuguese language is spoken by approximately 240 million of people around the
world and the number of published articles by Brazilian researchers and Brazilian
researcher’s citations has been progressively increased[18].
The MNSI-Brazil version did not need specific transcultural changes in the application
forms; however few adaptations in the text from the introduction (how to use) had
to be made to improve the Brazilian Portuguese comprehension. The activities demanding
in the questionnaire are quite similar between English and Portuguese as native language
people. The Brazilian raters affirmed that the MNSI is a simple, fast and easy-to-apply
test performing it in around three minutes. The Brazilian experience confirme the
point of view from the specialists who used the original version of the MNSI[9],[19],[20].
We tested the MNSI-Brazil in a sample composed by patients who schooling at least
7 years and were able to read and write well. This reflected an efficiency to perform
the self-administered part of the MNSI-Brazil that was observed in a maximum of the
3 minutes to respond. Probablely, less years of schooling could change the time to
perform it; however we do not have a sample with participants who schooling less than
7 years to confirm this hypothesis.
Studies recording electrical activity in the peripheral nerve are considered the gold-standard
method to detect and diagnose the DSPN, even regarding their technical characteristics
to evaluate the nervous fibers function in an accurate and precise way. Although electrophysiological
studies be the best option to diagnose DSPN, they are expensive and, for this reason,
must not be the initial tool to screening; nerve conduction studies must be recommended
after the clinical approach, by which the MSNI-Brazil can provide additional information
to the therapeutical decision-taking[9],[10],[11],[19].
The MNSI-Brazil reveals that our sample presented few signals to conclude a neuropathy
diagnosis as observed in our results showing just two subjects scored near from the
minimum to be considered a suggestion of DSPN. Although the reliability had demonstrated
excellent repeatability, we had a sample without patients presenting several suggestion
of DSPN, once we had recruited them from healthy services which were giving care to
prevent DSPN.
The ICC values showed high inter-rater (ICC = 0.840) and test-retest (ICC = 0.864)
reliabilities which were considered excellent[21], showing an agreement between measures and supporting the use of the MNSI-Brazil
for the evaluation of patients with diabete mellitus ([Figure]). The bland-altman plots ([Figure]) allows us to observe a magnitude of error absent (zero), very close to zero or
no more than 2 points for inter-rater and test-retest analysis. Psychometrical studies
pointing to be understandable to find a small error caused by sources issued by performance,
concentration, learning, distraction and others[22],[23].
At the end, the results showed evidence to support the use of the MNSI-Brazil with
repeated measures acceptable to suggest this instrument among the health professionals
dealing with the DSPN in Brazil.
In conclusion, the Brazilian version of Michigan Neuropathy Screening Instrument (MNSI-Brazil)
did not show great semantics, linguistics ou cultural discrepances or diferences which
could suggest any restraints for Brazilian people with diabetes mellitus. The translation
and cross-cultural adaptation process to Portuguese language obtained success following
the methodological norms suggested and internationally accepted. Excellent reliability
and internal consistency were found, making the MNSI-Brazil a useful instrument to
evaluate the peripheral neuropathy signs and symptoms in Brazilians with diabetes
mellitus.
