The Science of Diabetes Self-Management and Care2023, Vol. 49(5) 362–373© The Author(s) 2023Article reuse guidelines:sagepub.com/journals-permissionsDOI: 10.1177/26350106231192354journals.sagepub.com/home/tde
Purpose: The purpose of the study was to develop an Italian version of the Revised Brief Diabetes Knowledge Test (DKT2), providing a cultural and linguistic validation supported by psychometrics and hypotheses testing.
Methods: This multimethods study was divided into 4 phases: (a) cultural-linguistic validation, with a translation and back-translation process; (b) confirmatory factor analysis (CFA) considering the original scale’s structure (knowledge and insulin-specific knowledge); (c) criterion validity via hypotheses testing; and (d) cross-group measurement invariance. The internal consistency reliability was assessed by the Kuder-Richardson Formula 20 (KR-20) of the overall scale.
Results: A total of 251 patients and 251 caregivers were enrolled. The CFA showed good goodness of fit for both patients and caregivers. The tested hypotheses supported criterion validity in both groups. Reliability was adequate: All KR-20 values in both groups and domains were higher than 0.60. The mean percentage of knowledge score on DKT2 was lower for patients than caregivers.
Conclusion: The DKT2 is a valid and reliable scale to assess overall knowledge of diabetes, considering its role in promoting appropriate self-care behaviors in patients with type 2 diabetes mellitus. The Italian version of DKT2 demonstrated reliability and validity, and it can be used by researchers and diabetes care and education specialists to assess a patient’s or population’s overall knowledge of diabetes.
Adults with diabetes are estimated to be 537 million worldwide.1 Diabetes prevalence has more than tripled since 2000, and it is predicted that it will further increase by 11.3% in the worldwide population by 2030, affecting 643 million people.1 Diabetes is recognized as one of the most serious and growing challenges to public health. Individuals who have diabetes face significant challenges that affect them and their families.2 These challenges can include severe and potentially fatal complications that can significantly reduce their overall quality of life and even lead to early death.
Adequate self-care of diabetes is relevant to face this health challenge,2,3 and patient education is the cornerstone of care.4 Accordingly, knowledge of diabetes is an important component of diabetes self-care.5 Diabetesrelated knowledge of patients and their caregivers is key for informed decision-making processes about diet, exercise, weight control, blood glucose monitoring, use of medications, foot and eye care, and control of macrovascular risk factors.3 Higher levels of diabetes-related knowledge seem to support participation in diabetes education programs and dietician visits, facilitating a good self-monitoring of blood glucose levels6,7 and adequate level of glycated hemoglobin (A1C).8 Furthermore, diabetes-related knowledge could indirectly impact patients’ selfefficacy and their beliefs in treatment effectiveness, consequently affecting medications and self-care.5
Assessing diabetes-related knowledge of patients with diabetes and their caregivers is strategic to develop effective interventions for those patients.5 To do that, valid and reliable tools to measure diabetes-related knowledge are needed.9 The review by Eigenmann et al10 critically described 3 measures for assessing diabetes-related knowledge: the Diabetes Knowledge Test (DKT), which was developed in 199811; the Diabetes Knowledge scales, developed in 198412; and the Audit of Diabetes Knowledge, developed in 2001.13 Among these, the DKT is the most widely utilized assessing instrument and has been used for longer than 2 decades to evaluate diabetes-related knowledge.14 In 2016, the DKT was modified by simplifying expressions, improving grammar, and making changes to meet current national standards, while the items’ numbers and topics were maintained. For example, the sentence “If you have taken intermediate-acting insulin (NPH or Lente), you are most likely to have an insulin reaction in . . .” was changed to “If you have taken rapid-acting insulin, you are most likely to have a low blood glucose reaction in . . . .”15 This new tool was then named Revised Diabetes Knowledge Test (DKT2)11,15,16 and was composed of 2 sections, each scored separately: the general knowledge subscale, consisting of 14 items, and the insulin-use specific knowledge subscale, consisting of 9 items only for individuals using insulin.
The DKT2 is available for the English-speaking population, and it was translated and validated into Arabic,17 but it is not available in the Italian context, undermining the possibility of clinicians assessing the diabetes-related knowledge of Italian patients with diabetes and their caregivers. Considering that in Italy the management of diabetes requires roughly 7% to 9% of the resources available for financing the national health system,18 the routine utilization of DKT2 for assessing diabetes-related knowledge of patients with diabetes and their caregivers could be pivotal for optimizing care plans and facilitating a tailored approach to design educational strategies.
For this reason, this study’s general purpose was to develop the Italian version of DKT2, providing a cultural and linguistic validation supported by psychometric evidence to confirm its dimensionality and criterion validity in patients with type 2 diabetes (T2DM) and their caregivers. Therefore, the specific aims were (a) to provide cultural-linguistic validation, with a translation and back-translation process; (b) to confirm the scale’s theoretical dimensionality and its internal consistency reliability; (c) to verify the criterion validity; and (d) to assess crossgroup measurement invariance.
Design. A multicenter cross-sectional study was conducted in 4 outpatient diabetes clinics in the north of Italy. People with T2DM and their informal caregivers were enrolled between July 2018 and October 2019. Signed informed consent was obtained from all recruited participants. Approval was provided by the Institutional Review Board of each involved center. The study was conducted according to the ethical standards of the responsible committee on human experimentation and the Declaration of Helsinki.19
Participants. A consecutive sample of 251 patient-caregiver couples was recruited during outpatient visits at 4 outpatient diabetes clinics in the north of Italy. Inclusion criteria for patients were diagnosis of T2DM according to guidelines criteria,20 age ≥ 18 years, and presence of written informed consent. Exclusion criteria for patients were inability to read, time since diagnosis of T2DM <1 year, first access to the diabetes center, and documented cognitive impairment. Inclusion criteria for caregivers were being the main informal caregiver of the patient, age ≥ 18 years, and presence of written informed consent. Exclusion criteria for caregivers were inability to read and documented cognitive impairment.
Sample size. The sample size was calculated to ensure proper factor analysis so that the moderate factor structure ratio between subjects and items was 10:1.21 Considering the 23-items scale, thus, required a minimum of 230 respondents per group. Considering the 14-item and 9-item subscales, thus, required a minimum of 140 and 90 respondents, respectively, per group.
Measurement. Patients’ characteristics were collected through a data collection form. Data on age (years), sex (male, female), marital status (married/cohabitant, separated/divorced, widower/widow, single/never married), school education (primary school, secondary school, high school, university), employment status (employed, retired, unemployed, housewife), and cohabitation with the caregiver (yes, no) were included in the case report forms for patients. The same data were collected for caregivers, with additional information about the relationship with the patient (son/daughter, partner, brother/sister, son/daughter-in-law, other), caregiving hours per day (hours/day), years involved in diabetes caring (years), and participation in formal and structured events on diabetes (yes, no). Patients’ clinical data were the time from T2DM diagnosis (years), presence of T2DM complications (yes, no), BMI (kg/m2 ), last available value of A1C (%), and insulin use and were collected by reviewing medical records. Finally, to reduce socially desirable response bias, participants were allowed to answer the following questionnaires alone and without the support of the investigators.22
The DKT2 is a self-report tool used for assessing diabetes-related knowledge, and it can be used by both patients and caregivers.16 The DKT2 consists of 2 sections, each scored separately. The first section is composed of 14 items assessing general knowledge of diabetes regarding diet, glycemia, feet, physical activity, symptoms, and complications. The second section is an additional part of 9 items to be completed only in the presence of insulin therapy because this part investigates the knowledge about insulin therapy management. The score is the percentage of correct answers for each section.16 In its original language version, DKT2 validity and reliability have been demonstrated in people with diabetes.16 Particularly, all the defined a priori hypotheses for criterion validity were verified. In the present study, the DKT2 was administered to both patients and caregivers to evaluate the validity and reliability of the tool in each of the 2 groups.
Patient self-care self-efficacy and caregiver self-efficacy in contributing to patient self-care were measured administering the self-efficacy scale of the Self-Care of Diabetes Inventory (SCODI) and Caregiver Contribution to Self-Care of Diabetes Inventory (CC-SCODI), respectively.23 According to the middle range theory of self-care of chronic illness,23 the self-care self-efficacy scale assesses the self-confidence in performing self-care and in persisting in those behaviors despite barriers.23,24 This scale is composed of 11 items: Each item uses a 5-point Likerttype scale from “never” to “always” and provides a 0 to 100 standardized score where a higher score means better self-care self-efficacy.24 The scale of caregiver contribution to self-care self-efficacy uses the same scoring system of the previous one and differs only for the question asked assessing the caregiver confidence and persistence in supporting the patient in self-care: “In reference to the person you care for, how much do you feel confident that you can recommend or do these activities?”25
DKT2 back-translation. To achieve the Italian version of the DKT2, among the several possible translation techniques, a back-translation process was adopted.26 Back-translation involves a team of independent translators and aims at maintaining equivalence between original and translated versions.27 The process requires a first translator rendering the instrument from the original language to the target language and another translator independently back-translating it to the original language. A meticulous comparison ensures concept equivalence. Any mistakes or inconsistencies require a new translation.27
For the DKT2, first, 2 Italian nurses fluent in English and with experience in the field of diabetes research provided the translation of the DKT2, preserving semantic equivalence and accounting for cultural nuances. Second, a bilingual researcher living in the United Kingdom and working in the health field but unfamiliar with DKT2 retranslated the Italian version into English independently. Lastly, this new English version of the DKT2 was submitted to the authors of the DKT2 for their approval. Because the 2 English versions were conceptually equivalent and did not have errors in meaning,26,27 the Italian translation of the DKT2 was found to be adequate.
Hypotheses for criterion-related validity. When testing criterion-related validity, there are 2 types of hypotheses: concurrent and predictive. Concurrent validity hypotheses are tested when the criterion measure and the test measure are obtained at the same time. Considering the cross-sectional nature of data collection, this study tested criterion-related validity in the framework of concurrent validity.
Precisely, test scores were examined by level of formal education, operationalized as lower than secondary school and from high school onward, and by age, operationalized as participants ages <60 years versus participants ages ≥60 years. Following the original validation article16 and in accordance with previous research on T2DM,28-30 it was hypothesized that participants with more advanced education would score higher than participants with lower formal education. Moreover, considering the already known association between age and knowledge,28,29,31 it was hypothesized that younger participants would report higher scores than the ones reported by older participants.
DKT2 scores were also examined in relation to family support, operationalized as the cohabitation of another person who attends to the patient’s needs (caregiver). Considering the previously described association between family support and patient’s knowledge,32,33 it was hypothesized that patients with family support would score higher than patients without family support.
DKT2 scores of caregivers were also examined by years of caregiving experience, operationalized as ≥10 years of experience versus <10 years of experience. Considering the association between experience and knowledge in other populations,34 it was hypothesized that caregivers for whom ≥10 years have passed since they started to support patients with caregiving would score higher than caregivers for whom <10 years have passed since they started caregiving.
Considering the crucial role of self-efficacy as a mediator between knowledge and self-care in adults with heart failure,35 a significative positive correlation was hypothesized between knowledge and self-care self-efficacy (SCODI and CC-SCODI subscales).
Data analysis. Continuous sociodemographic and clinical variables, not being normally distributed, were described by median and 1st to 3rd quartile, respectively, as measures of central tendency and dispersion. Qualitative sociodemographic and clinical variables were described using absolute frequency and percentage. The psychometric properties of the DKT2 were then established by examining internal consistency reliability, structural validity, criterion validity, and measurement invariance.
To determine the internal consistency reliability of the DKT2, the Kuder-Richardson Formula 20 (KR-20) was adopted. KR-20 is used specifically for dichotomous data, meaning data with only 2 possible outcomes: correct versus wrong answers in relation to DKT2. The KR-20 is calculated based on the proportion of correct responses and the number of items on the test. The KR-20 ranges from 0 to 1, with higher values indicating a higher level of internal consistency. Generally, a KR-20 of 0.6 or higher is considered an adequate internal consistency level, indicating that the items on the test or scale measure a similar construct.36
A confirmatory factor analysis (CFA) was conducted separately per group to establish a valid model prior to the addition of covariates and confirm the theoretical dimensionality of the DKT2. The weighted least square mean and variance adjusted estimation method was used for fitting the CFA models to dichotomous data. It is a robust estimation method that can handle nonnormality and nonindependence of errors, which are common in dichotomous data.37 Multiple indicators multiple causes (MIMIC) modeling38 was used to test the effects of the covariate on the 2-factor model, considering the original dimensionality (general knowledge and insulin-specific knowledge). The MIMIC model was employed on the overall sample (patients or caregivers), and a covariate indicating the specific group (patients or caregivers) was added to assess the cross-group measurement invariance. Finally, nonparametric tests and Spearman’s correlation coefficient were used to assess the criterion validity via hypotheses testing to investigate criterion validity. Statistical significance was set at P < .05. Statistical analyses were performed in IBM SPSS Statistics for Windows, Version 28.0 (Armonk, NY: IBM Corp) and Mplus 8.1 (Los Angeles, CA: Muthén & Muthén).
A total of 251 patients with T2DM and 251 caregivers were enrolled, among which 105 patients were treated with insulin. Patients showed a median (1st-3rd quartile) A1C level of 7.1% or 54 mmol/mol (6.6%-7.8% or 49-62 mmol/mol), with a median (1st-3rd quartile) year from diagnosis of 12 (6-20) years. The median (1st-3rd quartile) age of patients was 72 (67-79) years, and 55% were male, while the median (1st-3rd quartile) age of caregivers was 64 (54-71) years, and 29% were male. Most of the caregivers were the partner of the patients (66%), and the median (1st-3rd quartile) of the reported hours dedicated to caregiving was 3 (1-6) hours per day (Table 1). Nearly all the patients and caregivers were married or partnered (76% and 90%, respectively), and only 35% of the patients and 11% of the caregivers received diabetes-specific education.
Item 4 (“Which of the following is a ‘free food?’”) was removed because its content was considered “not consistent” with current diabetes care39 and education40,41 guidelines. Indeed, item 4 in both groups (patients and caregivers) resulted in the lowest correct percentage (patients = 18.4%; caregivers = 23.9%) and the lowest item-total correlation (patients = 0.07; caregivers = −0.05).
Therefore, considering the newly revised Italian version of the DKT2 with 13 items in the general knowledge section instead of 14 items in the original English version of the DKT2, median patient scores were lower than the caregiver ones. Among patients, the median and interquartile ranges (1st-3rd quartile) of the general score was 61.5% (46.1%-71.1%), and for the insulin use section, it was 55.6% (33.3%-77.8%). Among caregivers, the median (1st-3rd quartile) of the general score was 69.2% (53.8%-76.9%), and for the insulin use section, it was 55.6% (44.4% to 77.8%; Table 1).
In the Italian version of the DKT2, the first section is composed of 13 items assessing general knowledge of diabetes (about diet, glycemia, feet, physical activity, symptoms, and complications), while the second section is an additional part of 9 items to be completed only in the presence of insulin therapy because this part investigates the knowledge about insulin therapy management. For this reason, a 2-factor confirmatory model was specified separately in both groups (patients and caregivers) that fitted well to sample statistics. Precisely, in patients: χ2 (229) = 234.99, P = .379; χ2/df = 1.19; root mean square error of approximation (RMSEA) = 0.018; 90% CI, 0.001-0.028; comparative fit index (CFI) = 0.989; Tucker-Lewis index (TLI) = 0.988; weighted root mean square residual (WRMR) = 0.835. In caregivers: χ2 (206) = 249.79, P = .020; χ2/df = 1.21; RMSEA = 0.029; 90% CI, 0.012-0.041; CFI = 0.912; TLI = 0.902; WRMR = 0.948. In both groups, factor loadings were positive, indicating a positive relationship of the latent variable “general knowledge” with the first 13 observed measures and “insulin use knowledge” with the latter 9 measures, with explained variance per item always ≥10% (P ≤ .001). Moderate/high positive correlations between the 2 factors were observed in both groups (rpatients = .80, P < .001; rcaregivers = .55, P < .001).
The MIMIC model performed on the overall sample explained sample statistics as well: χ2 (228) = 320.420, P = .0001; χ2/df = 1.40; RMSEA = 0.028; 90% CI, 0.021-0.035; CFI = 0.921; TLI = 0.912; WRMR = 1.031. Factor loadings were positive, indicating a positive relationship of the latent variable “general knowledge” with the first 13 observed measures and “insulin use knowledge” with the latter 9 measures, with explained variance per item always ≥10% (P ≤ .001). No significant relationship was detected between the covariate “group” (patients vs. caregivers) and the latent variables and observed items, sustaining a measurement invariance.
The tested hypotheses were confirmed to support the validity of the Italian version of the DKT2 in both groups (Table 2). Patients with an educational level lower than high school graduate had significantly lower scores in both general (U = 3622.0, P < .001) and insulin use knowledge (U = 661.0, P = .005); patients with family support had significantly higher scores in general knowledge (U = 4803.5, P = .006) but not significantly in insulin use section (U = 811.0, P = .062). Patients younger than 60 years had significantly higher scores in general (U = 1834.0, P < .001) and insulin use knowledge (U = 359.5, P = .022). Finally, there was a positive correlation between selfcare self-efficacy, general knowledge (P = .001), and insulin use knowledge (P = .010).
Caregivers with an educational level higher than high school graduate (U = 3622.0, P < .001), with ≥10 years involved in diabetes caring (U = 5698.0, P = .003), and younger than 60 years (U = 5083.5, P = .001) had significantly higher scores regarding general knowledge, but there were not any significant differences in the insulin use section (U = 1118.5, P = .178). Finally, there was a positive correlation between self-care self-efficacy and both general knowledge (r = .19, P = .002) and insulin use section (r = .36, P < .001).
As shown in Table 3, the reliability assessment for the 2 groups and the 2 subscales was performed separately. Among patients, KR-20 values were 0.635 and 0.733 in general and insulin-specific knowledge, respectively. Among caregivers, KR-20 values were 0.639 and 0.711 in general and insulin-specific knowledge, respectively.
This study was aimed to develop the Italian version of DKT2, providing a cultural and linguistic validation supported by psychometric evidence among both patients with T2DM and their caregivers. Results confirmed the DKT2 dimensionality and its validity, reliability, and invariance across the 2 groups, which were characterized by a lower percentage of correct answers in general and insulin use knowledge sections compared with the original validation article.16
The low level of diabetes knowledge that emerged might be a reflex of several factors, such as limited access to health education,2,41 limited health literacy,42 low socioeconomic status,43 psychological factors (ie, anxiety and depression),44 and lack of engagement in self-care behaviors.24,45,46 Although this study was not designed to determine factors determining limited knowledge, the evidence sustaining validity, reliability, and measurement invariance in patients and caregivers with these specific characteristics related to a relatively moderate/low knowledge is a novelty in the broader landscape of literature sustaining the adoption of DKT2 for clinical and research purposes.
In this study, only 1 item (i.e., Item 4, “Which of the following is a ‘free food?’”) was removed after a preliminary analysis because its content did not reflect current recommendations regarding diabetes-related nutrition,39,41 in which it is not reported that there are “free food.” In fact, people with diabetes can eat a variety of foods without cutting sugars and fats from the diet but staying mindful of their overall carbohydrate and calorie intake and the types of carbohydrates they consume.39 In general, people with diabetes should aim to consume a diet that is rich in fruits, vegetables, whole grains, lean protein, and healthy fats.39 In addition to the content aspects, item 4 showed the lowest item-total correlations in both groups (patients = 0.07; caregivers = −0.05), and a “floor effect” (ie, the situation in which a large proportion of participants perform very poorly on a task, thus skewing the distribution of scores and making it impossible to differentiate among the many individuals at that low level) was detected because almost all patients and caregivers answered with a wrong choice.47 This result may have occurred because the item can be considered “not consistent” with current diabetes care education plans.41
The 2-factor CFAs confirmed the originally hypothesized dimensionality of the DKT2 with 28 items in patients, caregivers, and in the overall sample with the MIMIC. In addition to confirming the scale dimensionality, this is the first study describing the measurement invariance of the DKT2 scores between patients and caregivers. Measurement invariance refers to the extent to which the measurement properties of a test or measure are equivalent across groups or over time.38,48 In relation to DKT2, evidence of measurement invariance allows researchers and clinicians to perform valid group comparisons because it helps determine whether differences in test scores across groups are due to actual differences in the construct being measured or due to differences in the measurement properties of the test or measure. This evidence could support future dyadic research involving patients and caregivers concerning their diabetes-related knowledge. However, these results should be verified in a future study using larger samples in different countries to increase the external validity of the available findings and assess the other levels of measurement invariance (ie, metric, scalar, and residual invariance).48
The hypotheses testing confirmed the criterion-related validity of the DKT2 in the Italian context. As expected, in both groups, participants with higher diabetes-specific knowledge are likely to have an educational level higher than high school,28-30 to be younger than 60 years old,28,29,31 and have higher self-care self-efficacy.35 Moreover, as per previous literature,32,33 patients with family support are likely to have higher diabetes-specific knowledge than their counterparts. Finally, considering caregivers, participants with higher diabetes-specific knowledge are likely to have more than 10 years of experience in diabetes caring,34 supporting known-groups validity.
In addition, the internal consistency of this 28-item DKT2 Italian version had adequate internal consistency and reliability. Specifically, the KR-20 provided similar results to the internal consistency shown in previous studies.11,15,16 Previous studies employed Cronbach’s alpha to calculate the internal consistency reliability of the DKT2. Considering that KR-20 is specifically designed for dichotomous data, its utilization for assessing the internal consistency of the DKT2 could have advantages in its accuracy because it is less sensitive to the number of items in the scale and, therefore, less likely to inflate the reliability coefficient when the number of items in the scale is increased, as per the case of the general knowledge of DKT2 (14 items). In addition, KR-20 is less affected by the correlation between items than Cronbach’s alpha, which means that it is less affected by highly correlated items.
The results of this study require caution in their generalizability because of the employed cross-sectional design. Therefore, this study did not provide any information regarding the stability of the validated 28-item Italian version of the DKT2. Furthermore, considering the limited sample size, it was not possible to perform additional subgroup analyses to provide more evidence of measurement invariance between specific subgroups.
Assessing diabetes-related knowledge is key to identifying when a patient with diabetes or a caregiver may have low levels of knowledge, which can then be addressed through education and support, acknowledging that knowledge can influence behaviors of self-care, such as adherence to treatment, and improve communication between patients, caregivers, and health care providers. The Italian version of DKT2, encompassing 22 items and measuring the general diabetes-related knowledge and the specific knowledge regarding insulin use, showed good evidence of reliability and validity in patients and caregivers.
The DKT2 scale is a tool that can be used to assess diabetes-related knowledge in Italian patients and, with this study, also in caregivers. Having evidence of validity and reliability regarding this tool in the specific Italian context is important because the emerging evidence sustains a proper utilization of this tool to identify lacks in diabetesrelated knowledge that a patient with diabetes or a caregiver may have and could be used to evaluate the effectiveness of diabetes education programs. More precisely, the DKT2 assesses knowledge in several areas, such as general diabetes information, diabetes complications, and diabetes self-care. By identifying areas where patients lack knowledge, health care providers can target education and resources to those specific areas. Assessing diabetes-related knowledge can be also relevant as a measure for quality improvement: By assessing the level of knowledge and providing the necessary education and support to improve it, the DKT2 could be used as a tool for quality improvement in health care settings.
Valid and reliable measurements of diabetes-related knowledge are pivotal to provide accurate information about a patient’s or caregiver’s understanding of their condition, which can inform health care providers about areas where additional education and resources are needed. The evidence of validity and reliability in specific settings, such as the Italian context in this study, is also important to make more informed treatment decisions and develop a tailored care plan that addresses the patient’s specific needs and lacks in knowledge levels. Broadly, valid and reliable measurements of diabetes-related knowledge can inform health care policy by providing accurate information about the level of knowledge in a population (population-based studies) and by evaluating the effectiveness of interventions aimed at increasing knowledge.
This study’s abstract was presented in part at the International Council of Nurses (ICN 2021) Congress, November 2-4, 2021 (virtual).
The authors declare that there is no conflict of interest.
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
This study was conducted in accordance with the ethical principles that have their origin in the Declaration of Helsinki and that are consistent with GCP and the applicable regulatory requirement(s). This study received approval from the local Institutional Review Board, and all participants provided written informed consent prior to participation.
Rosario Caruso https://orcid.org/0000-0002-7736-6209
Diletta Fabrizi https://orcid.org/0000-0002-5152-8804
Michela Luciani https://orcid.org/0000-0001-7598-5658
Davide Ausili https://orcid.org/0000-0001-5212-6463
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Supplemental material for this article is available online.
From Health Professions Research and Development Unit, IRCCS Policlinico San Donato, Italy (Ms Baroni, Dr Caruso); Department of Public Health, Experimental and Forensic Medicine, Section of Hygiene, University of Pavia, Pavia, Italy (Dr Arrigoni); Department of Medicine and Surgery, University of Milan-Bicocca, Monza, Italy (Mrs Fabrizi, Dr Pinto, Dr Michela, Dr Ausili); and ASST Grande Ospedale Metropolitano Niguarda, Milan, Italy (Dr Pinto).
Corresponding Author:Diletta Fabrizi, Department of Medicine and Surgery, University of Milan-Bicocca, Via Cadore 48, Monza, 20900, Italy.Email: d.fabrizi@campus.unimib.it
