Benazizi et al. Health Qual Life Outcomes (2021) 19:189 https://doi.org/10.1186/s12955-021-01818-5 RESEARCH Open Access Adaptation and psychometric validation of Diabetes Health Profile (DHP‑18) in patients with type 2 diabetes in Quito, Ecuador: a cross‑sectional study Ikram Benazizi1* , Mari Carmen Bernal‑Soriano1,2, Yolanda Pardo2,3,4, Aida Ribera2,5, Andrés Peralta‑Chiriboga1,6, Montserrat Ferrer2,3,4, Alfonso Alonso‑Jaquete7, Jordi Alonso2,3,8, Blanca Lumbreras1,2 and Lucy Anne Parker1,2 Abstract Introduction: The Diabetes Health Profile (DHP‐18), structured in three dimensions (psychological distress (PD), barri‑ ers to activity (BA) and disinhibited eating (DE)), assesses the psychological and behavioural burden of living with type 2 diabetes. The objectives were to adapt the DHP‐18 linguistically and culturally for use with patients with type 2 DM in Ecuador, and to evaluate its psychometric properties. Methods: Participants were recruited using purposive sampling through patient clubs at primary health centres in Quito, Ecuador. The DHP-18 validation consisted in the linguistic validation made by two Ecuadorian doctors and eight patient interviews. And in the psychometric validation, where participants provided clinical and sociodemo‑ graphic data and responded to the SF-12v2 health survey and the linguistically and culturally adapted version of the DHP-18. The original measurement model was evaluated with confirmatory factor analysis (CFA). Reliability was assessed through internal consistency using Cronbach’s alpha and test–retest reproducibility by administering DHP- 18 in a random subgroup of the participants two weeks after (n = 75) using intraclass correlation coefficient (ICC). Convergent validity was assessed by establishing previous hypotheses of the expected correlations with the SF12v2 using Spearman’s coefficient. Results: Firstly, the DHP-18 was linguistically and culturally adapted. Secondly, in the psychometric validation, we included 146 participants, 58.2% female, the mean age was 56.8 and 31% had diabetes complications. The CFA indi‑ cated a good fit to the original three factor model (χ2 (132) = 162.738, p < 0.001; CFI = 0.990; TLI = 0.989; SRMR = 0.086 and RMSEA = 0.040. The BA dimension showed the lowest standardized factorial loads (λ) (ranging from 0.21 to 0.77), while λ ranged from 0.57 to 0.89 and from 0.46 to 0.73, for the PD and DE dimensions respectively. Cronbach’s alphas were 0.81, 0.63 and 0.74 and ICCs 0.70, 0.57 and 0.62 for PD, BA and DE, respectively. Regarding convergent validity, we observed weaker correlations than expected between DHP-18 dimensions and SF-12v2 dimensions (r > −0.40 in two of three hypotheses). *Correspondence: ibenazizi@umh.es 1 Department of Public Health, Universidad Miguel Hernández, Sant Joan d’Alacant, Alicante, Spain Full list of author information is available at the end of the article © The Author(s) 2021. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/. The Creative Commons Public Domain Dedication waiver (http://​creat​iveco​ mmons.​org/​publi​cdoma​in/​zero/1.​0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Benazizi et al. Health Qual Life Outcomes (2021) 19:189 Page 2 of 12 Conclusions: The original three factor model showed good fit to the data. Although reliability parameters were adequate for PD and DE dimensions, the BA presented lower internal consistency and future analysis should verify the applicability and cultural equivalence of some of the items of this dimension to Ecuador. Keywords: Diabetes Health Profile-18, Diabetes mellitus, Quality of life, Validity, Reliability, Ecuador Background Although there is a wide description of different ques- Diabetes mellitus (DM) is a high priority public health tionnaires to assess quality of life in diabetic patients [18], problem. It is the most frequent chronic disease in the such as the Diabetes Care Profile, which aims at assess- world and, in 2014, affected 422 million people. Accord- ment of factors important in a patient’s adjustment to ing to the World Health Organization, people with type 2 diabetes and its treatment in daily life and which consists Diabetes mellitus (T2DM) represent 90% of all diabetics. of 234 items, the Appraisal of Diabetes Scale, which aims The prevalence of T2DM has increased more rapidly in at Assessment of diabetes-related distress and which low- and middle-income countries than in high-income consists of 7 items, the Diabetes Distress Scale, which countries, as is the case in Latin America and Ecuador aims to Measure of diabetes-related emotional distress [1]. In 2016, the prevalence of T2DM in Ecuador was for use in research and clinical practice and which con- estimated at 7.3% and has been rising significantly in all sists of 17 items, among others [19]. We chose the Diabe- age groups [2–5]. According to data from the STEPS Sur- tes Health Profile (DHP) because of its advantages over vey of Ecuador in 2018, the prevalence of diabetes was other diabetes-specific patient reported outcome meas- 6.6% in both sexes (6.6% in men and 6.5% in women) of ures. It is a specific instrument to evaluate psychologi- the Ecuadorian population between 18 and 69 years of cal and behavioural impact of living with diabetes [20]. age, and increased to 10.7% in the age group between 45 It generates a health profile that measures psychological and 69 years in both sexes [6]. distress, barriers to activity, and uninhibited eating. Each T2DM is the most common metabolic cause of mor- answer is rated on a scale, and the scores by dimension tality, due to its complications and associated patholo- are presented on a scale in which a higher DHP value is gies [7]. It negatively affects quality of life [8], defined as a associated with a worse perception of quality of life. The person’s individual perception of the physical, emotional short version of DHP with 18 items has been used in dif- and social state [9], as a result of associated physical dis- ferent countries, demonstrating adequate metric proper- abilities and mental health problems [10]. Clinical meas- ties [21–23]. ures can provide a good estimate of disease control, but The objectives of this study are to adapt the Diabetes the ultimate goal of DM care is to maintain or improve Health Profile-18 (DHP‐18) both linguistically and cul- the patient’s quality of life [11]. turally for use with patients with T2DM in Ecuador, and There are generic instruments to measure quality of to evaluate its psychometric properties. life that can be used both in the general population and in all disease groups [12, 13]. However, specific instru- Methods ments have been developed to measure specific effects Participants of diseases and are more responsive to changes. Disease- We included type 2 diabetic patients, who were at specific instruments can help determine which condi- least 18 years of age, had been diagnosed for at least tions best explain a patient’s limitations in physical and 12 months, resided in Quito with no intention of mov- / or mental function, and, therefore, are more useful in ing in the near future and were native Spanish speak- outcome research, health care cost studies, and clinical ers. Recruitment to the study used purposive sampling practice [14]. through a patient club for people with diabetes at the In Ecuador, advanced age, longer disease duration, Chimbacalle Health Center and contacts from health hypertension and kidney disease are associated with a promoters from several health centres in Quito (Número lower health related quality of life in patients with T2DM 1, Jardín del Valle, Cotocollao, Jaime Roldos Aguilera, [15, 16]. In addition, a direct relationship was found Corazón de Jesus, Comité del Pueblo, San Antonio de between low socioeconomic status and the development Pichincha, Colinas del Norte, Pomasqui, Carcelén Bajo, of the disease [17]. El Condado, Mena del Hierro, La Bota, Pisulí, Puellaro, Despite the rapid growth in the prevalence of T2DM Chavezpamba, Cotocollao Alto and Calacalí). and the existence of different instruments to measure In this setting, diabetic patient’s clubs are sometimes quality of life in diabetic patients, none of them have been established in primary health care centres, either by ini- linguistically or psychometrically validated in Ecuador. tiative of the health staff or the patient’s themselves. The Benazizi et al. Health Qual Life Outcomes (2021) 19:189 Page 3 of 12 role of patient clubs is to motivate patients through the do you assess your condition today? (1) unchanged, (2) exchange of experiences among its members, in addition improved, (3) greatly improved, (4) impaired or (5) highly to the orientation, advice and guidance offered by health impaired”. professionals on behaviour modification (physical activ- ity/diets) [24, 25]. Data collection Our selection sought to include a group of patients that The 8 interviews carried out during the linguistic and was heterogeneous in terms of sex, age and level of edu- cultural adaptation were held face to face but given the cation. All participants gave their consent to participate situation generated by the COVID19 pandemic [27], in the study. the data for the psychometric validation was collected through individual telephone interviews. Responses were Procedure digitally recorded by the interviewer using the Kobo tool- The interviews were carried out between February and box (http://​www.​kobot​oolbox.​org/) free open-source July 2020. The DHP-18 validation process consisted of 2 software on electronic tablets. Informed consents were phases. provided orally and were audio recorded. Linguistical and cultural adaptation DHP‑18 questionnaire Two Ecuadorian medical researchers reviewed the Participants responded to the adapted version of DHP- original version of the DHP-18 (English) and the exist- 18. We used the Diabetic Health Profile (DHP) -18 ing translation (Spanish for the United States) to assess because it is a shortened version of DHP-1, a specific the cultural and linguistic relevance for its use in Ecua- instrument for measuring the psychological and behav- dor. They suggested some changes in text, as well as the ioural impact of type 1 diabetes. We decided to use the reasons for these changes and provided a new recom- short version of the DHP because it can be used in people mended translation. Changes were discussed with the with both type 1 and type 2 diabetes aged 11 and older. other members of the team and a new adapted version And because the instrument has demonstrated adequate of the questionnaire was proposed. Subsequently, 2 dif- metric properties and its completion time is approxi- ferent researchers carried out interviews to assess the mately 5–6 min. Items are scored using a 4-point Likert- linguistic and cultural understanding of the adapted type scale ranging from 0 to 3. Items are provided with questionnaire with 8 people with T2DM of Ecuadorian one of four sets of responses (1) never, sometimes, gen- nationality in the Chimbacalle Health Centre. Partici- erally, always; (2) never, sometimes, often, very often; pants were asked to answer the questions and then, the (3) not at all, a little, a lot, very much; and (4) very likely, necessary time was recorded, the answer options were quite likely, unlikely, not at all likely. The raw subscale discussed, the wording that was difficult to understand scores are transformed into a common score range from was commented, and alternative wording was suggested 0 to 100, with 0 representing no dysfunction. based on the participants’ own words. A second adapted The DHP-18 consists of three dimensions: psycho- version was proposed. The interviews were recorded and logical distress (includes questions like depressed from transcribed verbatim for analysis. Finally, participants’ diabetes; more arguments or upsets at home than there responses were summarized in a pilot test report includ- would be if you did not have diabetes; losing your tem- ing recommended changes and suggestions. The report per over unimportant things; etc.), barriers to activity was then sent to the original authors of the questionnaire (includes questions like food controls life; difficult stay- for verification and approval. ing out late; avoid going out when sugar is low; etc.) and disinhibited eating (includes questions hard to say no to Psychometric validation food you like; ease of stopping when you eat; wish there Firstly, we recruited 146 participants for the baseline test were not so many nice things to eat; etc.). where they responded to the questions posed in the tool previously linguistically validated DHP-18 instrument in SF‑12 v2 Ecuador and in another tool (SF-12v2 in its version for The SF-12 v2 is an instrument for measuring health- use in Ecuador) [26] in order to assess the correlation related quality of life [26], based on SF-36. It includes with generic quality of life as a construct validity test. twelve items, has an application time of approximately Two weeks later, we assessed the intra-observer reli- two minutes, and is used to evaluate the degree of well- ability of the new tool in a random sample of 75 of the being and functional capacity of people over 14 years of previously interviewed patients, where only DHP-18 was age. The response options form Likert-type scales (where retested, along with the following question: “Compared the number of options varies from three to six points, to the last time you completed the questionnaire, how depending on the item), which assess intensity and / or Benazizi et al. Health Qual Life Outcomes (2021) 19:189 Page 4 of 12 frequency of people’s health status. The score ranges from [32] and confirmed for other language translations [23]. 0 to 100, where the higher score implies a better health- In this case, we used CFA using Diagonally Weighted related quality of life. The SF-12v2 has demonstrated Least Squares (DWLS) [33–35] to test the hypothesis that adequate validity and reliability in the United States and the general construct of DHP is composed of three indi- internationally, and the Spanish version has been used vidual and correlated factors: psychological distress (6 successfully in Latin America and with Spanish-speaking items), activity barriers (7 items) and disinhibited eating populations in the United States. Investigations that use (5 items). To estimate the model fit, we used the following these twelve items of the SF have verified that the instru- criteria: Values > 0.95 for the Tucker-Lewis index (TLI) ment is a valid and reliable measure in Latin American or for the comparative fit index (CFI) and the root mean countries such as Colombia and Chile in adult popula- square error of approximation (RMSEA) < 0.06 or the tion, and a translated version is available for Ecuador. standardized root mean square residual (SRMR) < 0.08 The SF12v2 includes questions related to health status are considered as a good model fit [36, 37]. The magni- and limitations in doing activities, problems with work or tudes of factor loadings of 0.3 or greater were considered other regular daily activities due to physical health, due suitable. to emotional problems, pain, feelings, etc. Sociodemographic and clinical variables Reliability We collected sociodemographic and clinical variables (all To measure internal consistency reliability, we used self-reported by the participants): age, sex, marital status, Cronbach’s alpha coefficient, where values > 0.7 are con- ethnicity (mestizo or other minorities. The mestizos are sidered as acceptable [36]. The homogeneity of items was an ethnicity composed of Spanish and indigenous herit- verified by the analysis of item-rest and inter-item cor- ages), educational level, monthly income, employment relations for the items constituting each dimension of status, smoking status, alcohol intake, weight, height, the scale. The usual rule of thumb is that an item should duration of illness, use of medications, diabetes compli- correlate between 0.3 and 0.7 with the total score of the cations and comorbidities. factor (excluding that item), using Pearson’s coefficient. Additionally, average inter-item correlations for items Statistical analysis in the same factor should correlate moderately, between We included descriptive statistics through frequencies, 0.15 and 0.5, to ensure that they measure the same con- the mean (standard deviation) or the median (inter- struct but not so closely as to be too redundant [38]. quartile range), as appropriate. The psychometric char- We measured test–retest reliability in patients report- acteristics of the DHP-18 were assessed according to ing no-change in the global assessment of change ques- consensus-based standards for the selection of health tion. To measure test–retest reliability we considered status measurement instruments (COSMIN) guidelines that the individual’s health was significantly better if [28]. Missing values for the DHP-18 and SF-12 v2 were they responded, “much better” or “somewhat better” substituted with the mean of the completed questions for in the global assessment, or significantly worse if they those dimensions in which ≥ 50% of questions had been responded “somewhat worse” or “much worse” [39]. We completed [29, 30]. used the intraclass correlation coefficient (ICC) under We evaluated floor and ceiling effects by calculating a 2-way random effects model with absolute agreement the percentage of patients scoring either the lowest or [40], and its associated 95% confidence interval. We con- highest possible dimensional scores. If more than 15% of sidered that a questionnaire exhibits substantial reliabil- respondents achieve the lowest or highest possible score, ity when ICC is between 0.40 and 0.75, and greater than then floor or ceiling effects are present [31]. 0.90 represents excellent reliability [36]. Statistical analyses were performed using Stata Version Measurement errors were determined by calculating 15 (StataCorp LP; College Station, TX) and R software, the standard error of measurement (SEM) and the small- version R 4.0.0 (R Core Team. R: A language and environ- est detectable change (SDC). We calculated SEM by the ment for statistical computing. R Foundation for Statis- square root of the error variance derived from analysis tical Computing, Vienna, Austria; http://​www.R-​proje​ct.​ of variance (ANOVA), two-way ANOVA with repeated org) was used to perform the confirmatory factor analy- measures [41]. The ­SDCindividual and ­SDCgroup was calcu- sis. The level of statistical significance was set at p < 0.05. lated with the following formulas (41): √ SDCindividual = 1.96 ∗ 2 ∗ SEM Structural validity We performed a confirmatory factor analysis (CFA) SDCgroup = ­(SDCindividual /√n); n: number of subjects in because the factor structure had already been determined the sample. Benazizi et al. Health Qual Life Outcomes (2021) 19:189 Page 5 of 12 We estimated the minimally important difference H5: Individuals with a higher education level would (MID) for each DHP-18 dimension using three distribu- have lower DHP-18 values (better quality of life) than tion-based methods to estimate MID: 0.2 and 0.5 stand- those with a lower education level [47, 48]. ard deviation (SD) and SEM estimations. Formulas: H6: Individuals with diabetes-related complications would have higher DHP-18 values (poorer quality of life) 0.2SD = 0.2 ∗ SDbasaline than patients without complications [44]. 0.5SD = 0.5 ∗ SDbasaline 1SEM = SEM Results We also estimated Cohen’s d effect size (ES) of the Linguistic and cultural adaptation change in DHP-18 dimensions for those reporting a small Two Ecuadorian medical researchers modified some lin- but important change and those reporting no changes in guistic expressions in the Spanish version for the United global assessment rating. Cohen’s d was calculated with States in items 2, 3, 4, 5, 6, 10, 11, 12, 13, 14, 15 and 18 the following formula (42): and in some answer options. In the linguistic and cultural review, six women and two men participated: a 28-year- ES = (Scorebaseline − Scoreretest )/SDbasaline old person, a 49-year-old person and a 52-year-old per- son, and the rest of the participants were over 70 years SDbasaline: Standard deviation of baseline score. old. They made further changes to items 5, 6, 10 and An effect size of 0.2 was considered small, 0.5 moderate 12 and proposed reformulation of some expressions. and 0.8 large [43]. Most modifications were minor linguistical issues to use terms more commonly used in Ecuador, for example the expression “staying out” was changed by “going out of the Construct validity house”, the term “edgy” was changed by “nervous”, and We assessed construct validity of the DHP questionnaire the term "lose your temper" was changed by "get angry using three approaches. Firstly, we assessed convergent easily". Other changes were made to improve compre- validity using binary correlation analysis (Spearman’s r- hension by simplifying technical terms, for example due to non-normal value distributions) of the DHP-18 “influenza” was changed to “flu”, “depressed” was changed and SF-12v2. Before starting the analysis, we set up the to “sad”. One of the items was flagged as having potential following a priori hypothesis: (1) Scores of “psychologi- difficulties because participants would be asked to reflect cal distress” dimension in DHP-18 correlate negatively on their sugar levels, and there was a very low availabil- with scores of “mental health” dimension in SF-12v2. (2) ity of glucometers in homes. The expression "on the low Scores of “activity barriers” dimension in DHP-18 corre- side" was changed to "having low or very low sugar lev- lates negatively with “physical dimension” in SF-12v2. (3) els". Similarly, in item 6, the word "monitor" was replaced Scores of “disinhibited eating” dimension in DHP-18 cor- by the expression "take the sugar test" to improve its relates negatively with “physical dimension” in SF-12v2. understanding. The original author approved the new Secondly, we explored discriminant validity by compar- tool, linguistically and culturally adapted to the context ing the correlation among the three dimensions of the of Quito, Ecuador. DHP-18 scale. Thirdly, we evaluated known-group validity by compar- ing DHP-18 scores in patients according to sex, education Psychometric validation level, obesity, and clinical characteristics such as duration We recruited 146 patients diagnosed with T2DM. Table 1 of diabetes, presence of comorbidities and/or diabetes- describes the characteristics of the study population. The related complications using a Student’s t-test or ANOVA. mean age of the participants was 56.8 years, 58.2% were We tested the following pre-defined hypotheses: women and 80.1% were mestizo. The population studied H1: Individuals with longer duration of illness would had relatively low educational qualifications, with 56.8% have higher DHP-18 scores (poorer quality of life) than having primary or no education, 27.6% were not working those with shorter illness duration [44]. and 61.4% had incomes of less than $375 per month. H2: Obese individuals would have higher DHP-18 val- Regarding diabetes medication, the majority were on ues (poorer quality of life) than non-obese individuals oral antidiabetic therapy (66.2%), 11.7% of patients were [45]. treated with insulin, 2.1% with only diet and the rest H3: Women would report higher DHP-18 values (20%) were on combined therapy (oral + insulin). We (poorer quality of life) than men [46]. found that 37.5% were overweight and 29.5% were obese. H4: Individuals with comorbidities would have higher Of the 146 respondents, 135 (92.5%) answered the 18 DHP-18 values (poorer quality of life) [44]. items of DHP-18; 9 (6.2%) omitted one item and 2 (1.3%) omitted two or more items. The missing values were in Benazizi et al. Health Qual Life Outcomes (2021) 19:189 Page 6 of 12 Table 1 Description of study population Characteristics Total Re-test p valor n: 146 (%) n:75 (%) Age Mean (SD) 56.8 (11.3) 57.5 (10.3) 0.467 Sex Female 85 (58.2) 41 (54.7) 0.404 Male 61 (41.8) 34 (45.3) Ethnicity Mestizo 117 (80.1) 59 (78.7) 0.683 Other minorities 29 (19.9) 16 (21.3) Civil status Fig. 1 Confirmatory Factor Analysis performed based on the Married/cohabitant 99 (67.8) 55 (73.4) 0.127 polychoric correlation matrix, using diagonally weighted least squares (DWLS). CFI = 0.990. TLI = 0.989. RMSEA = 0.040 (90% confidence Separated/divorced 18 (12.3) 10 (13.3) interval: 0.011–0.059). SRMR = 0.086 Single/widow-widower 29 (19.9) 10 (13.3) Educational level No formal schooling 10 (6.8) 3 (4.0) 0.561 Basic education 73 (50.0) 38 (50.7) Seventy-five (51.4%) participants were retested for High school and non-university 47 (32.2) 25 (33.3) DHP-18. There were no differences in socio-demographic higher education or clinical characteristics between participants who were University studies 16 (11.0) 9 (12.0) retested and those who were not (Table 1). In the DHP- BMI 18 retest, there were two missing values in item 4 and 18.5–24.9 kg/m2 28 (23.3) 12 (18.7) 0.159 two in item 14, the distribution of the missing were 2 par- 25.0–29.9 kg/m2 45 (37.5) 22 (34.4) ticipants who did not answer one item and 1 participant ≥ 30 kg/m2 47 (39.2) 30 (46.9) did not answer two items. Not reported 26 11 Time since diagnosis with diabetes Structural validity <  = 5 years 54 (37.0) 28 (37.3) 0.245 CFA with values of χ2 (132) = 162.738, p < 0.001; 6–10 years 45 (30.8) 27 (36.0) CFI = 0.990; TLI = 0.989; SRMR = 0.086 and > 10 years 47 (32.2) 20 (26.7) RMSEA = 0.040 indicated a good fit to the data, except Diabetes complications for SRMR. The standardized factorial loads (λ) from each At least one 31 (21.2) 16 (21.3) 0.976 item on their respective factors were all statistically sig- Macrovascular complication* 5 (3.4) 3 (4.0) 0.694 nificant (p < 0.001) and ranged from 0.57 to 0.89, from Microvascular complication 28 (19.2) 14 (18.7) 0.872 0.21 to 0.77 and from 0.46 to 0.73 for psychological dis- Ocular 23 (15.8) 11 (7.5) 0.965 tress, barriers to activity and disinhibited eating, respec- Renal 7 (4.8) 3 (2.1) 0.802 tively. The covariance between the three latent variables Neuropathy 2 (1.4) 0 ranged from 0.54 to 0.90, with psychological distress Diabetic foot 2 (1.4) 1 (0.7) 0.941 and disinhibited eating presenting the highest covari- Comorbidities ance. And two items (questions 1 and 3) showed a λ At least one 84 (57.5) 44 (58.7) 0.776 value below 0.3, using a one-factor model (Fig. 1). When Hypertension 65 (44.5) 37 (49.3) 0.229 we repeated the analysis excluding these 2 items, we Depression 7 (4.8) 2 (2.7) 0.266 observed a significant improvement in all the indicators, Dyslipidemia 8 (5.5) 4 (5.3) 0.879 including the SRMR, which was the only one that showed Other comorbilidities 10 (6.8) 2 (2.7) 0.051 a value slightly higher than recommended (CFI = 0.997; NA Not available TLI = 0.996; RMSEA = 0.027 (90% confidence interval: *Includes: Ischemic Heart Disease, Cerebrovascular Disease, Heart Failure 0.000–0.052); SRMR = 0.078). Reliability items 4, 5, 7,12, 13 and 14. One item (question 14) of the Overall Cronbach’s alpha was 0.77 and dimensional DHP-18 version showed unbalanced responses with 75% alphas were 0.81, 0.63 and 0.74 for psychological distress, of respondents reporting never. barriers to activity and disinhibited eating, respectively. The three dimensions were in a suitable range (0.15–0.50) Benazizi et al. Health Qual Life Outcomes (2021) 19:189 Page 7 of 12 for average interitem correlation values which ranged associated to disinhibited eating were statistically sig- from 0.38 to 0.47, from 0.17 to 0.26 and from 0.32 to 0.41 nificant. For H3 and H6, the expected tendency of scores for psychological distress, barriers to activity and disin- for each dimension was obtained with higher scores in hibited eating, respectively. Item-rest correlation values women than men and in patients with diabetes-related ranged from 0.39 to 0.61 for disinhibited eating while complications than those without, but there were no sta- values ranged from 0.43 to 0.72 for psychological dis- tistically significant differences (Table 4). tress dimension, where item 17 showed the highest value Regarding hypotheses H1, H4 and H5, score patterns (0.72) and ranged from 0.07 to 0.47 for barriers to activ- were different from those expected. Individuals with ity, where item 1 showed the lowest value (0.07). longer duration of illness (H1) had lower scores reflect- When we repeated the analysis excluding question ing improved quality of life, although the differences were 1 which had an item-rest correlation value below 0.30 not statistically significant. Similarly, regarding educa- (value: 0.07) and a λ value < 0.3 in the barriers to activ- tional level (H5), scores did not show a clear tendency, ity dimension, the dimensional and overall Cronbach’s with the exception of lower scores for barriers to activ- alpha changed to 0.67 and 0.76, respectively. When ques- ity dimension with increasing educational level. Finally, tion 17 (item-rest correlation value slightly higher than there were no differences by presence of comorbidities 0.7) was excluded from psychological distress dimension, (H4) but we found differences between patients with or the dimensional and overall Cronbach’s alpha changed to without specific comorbidities such as hypertension and 0.75 and 0.76, respectively. depression. Having hypertension was associated with ICC values for a total of 75 retested participants better evaluation of two dimensions (psychological dis- (Table 2) were 0.70 (95%CI: 0.57, 0.80), 0.67 (95%CI 0.56– tress and disinhibited eating), while depression was asso- 0.77), 0.73 (95%CI 0.64–0.81) for psychological distress, ciated with worse evaluation of two dimensions (barriers barriers to activity and disinhibited eating, respectively. to activity and disinhibited eating) (Table 4). Among the retest participants, thirty-nine (52%) reported that their condition was unchanged from baseline to retest (ICC values in Table 2) and 36 (48%) Discussion reported that their condition had changed from baseline In the present study, we linguistically and culturally to retest. Fifteen (20%) participants reported that their adapted the DHP-18 and investigated its psychometric condition had improved and 21 (28%) reported that their properties in people resident in Quito, Ecuador. Satisfac- condition had deteriorated. ICC values for participants tory psychometric properties were observed in a substan- reporting that their condition stayed the same were 0.69 tial number of aspects. The factor structure was adequate (95%CI 0.48–0.83), 0.66 (95%CI 0.50–0.79), 0.66 (95%CI but with 2 items, belonging to the dimension of barriers 0.50–0.80) for psychological distress, barriers to activity to activity, which were loaded below the recommended and disinhibited eating, respectively. value. Although reliability parameters were adequate for psychological distress and disinhibited eating dimen- Construct validity sions, the barriers to activity presented lower internal Our assessment of convergent validity showed an inverse consistency and future analysis should verify the appli- relationship between DHP-18 dimensions and SF12v2 cability and cultural equivalence of some of the items of dimensions and the results verified two of three a priori this dimension to Ecuador. hypotheses with correlation values between 0.4 and 0.7 Except for the dimension of barriers to activity, a good (Table 3). internal consistency was found. The internal consist- For discriminant validity, correlations between the ency of the dimension of barriers to activity contrasts DHP-18 dimensions were 0.4 or more, ranging from 0.40 with another study [21] and may be related to the differ- to 0.74. The highest correlation was between psychologi- ent populations of patients investigated [21, 22, 32], since cal distress and disinhibited eating (r = 0.74), followed by in some studies of people with both type 1 or 2 diabetes the correlation between psychological distress and barri- were included. Based on a more detailed analysis of the ers to activity (r = 0.45) and the lowest was the correla- total item statistics, we observed that the elimination tion between barriers to activity and disinhibited eating of items 1 and 17, with the lowest and highest item-rest (r = 0.40). correlation values, did not produce significant increases With regard to known-group validity, our results in overall and dimensional consistency, as observed in showed the expected tendency in three (H2, H3 and H6) another study [23]. of the 6 initial hypotheses. Compared to individuals with The test–retest reliability showed substantial reliabil- BMI < 30 kg/m2, those with BMI ≥ 30 kg/m2 (H2) showed ity values in accordance with the recommendations of higher values for each dimension, although only those the literature [36]. And the sample size used is within the Table 2 Test–retest reliability of the Diabetes Health Profile-18 subdimensions overall, and considering a reported change in global assessment of health Benazizi et al. Health Qual Life Outcomes Global assessment N Baseline, mean (sd) Retest, mean (sd) Mean differences (sd) ES Minimally important difference SDCind SDCgru ICC (95%CI) of health 0.2SD 0.5SD 1SEM PD (2021) 19:189 Stayed the same 39 22.4 (16.3) 21.5 (13.7) 0.8 (11.9) 0.05 3.27 8.17 8.43 23.37 2.65 0.69 (0.48–0.83) Worse 15 28.9 (17.7) 26.3 (17.5) 2.6 (10.9) 0.14 3.54 8.86 7.70 21.33 3.89 0.81 (0.53–0.93) Better 21 33.9 (23.2) 31.7 (24.8) 2.1 (20.6) 0.09 4.65 11.61 14.56 40.37 6.23 0.64 (0.30–0.84) Total 75 26.9 (19.2) 25.3 (18.5) 1.5 (14.5) 0.08 3.84 9.59 10.26 28.45 2.32 0.70 (0.57–0.80) BA Stayed the same 39 35.2 (15.7) 37.8 (20.4) − 2.6 (17.6) 0.17 3.14 7.84 10.01 27.75 2.59 0.66 (0.50–0.79) Worse 15 49.3 (13.2) 56.7 (19.2) − 7.4 (19.3) 0.56 2.63 6.58 11.38 31.55 4.81 0.22 (− 0.06,0.58) Better 21 40.1 (20.1) 43.5 (24.3) − 3.4 (20.1) 0.17 4.01 10.03 11.60 32.16 4.05 0.71 (0.51,0.86) Total 75 39.4 (17.2) 43.2 (22.3) − 3.8 (18.3) 0.22 3.45 8.61 12.67 29.57 1.99 0.67 (0.56–0.77) DE Stayed the same 39 30.0 (19.0) 32.1 (21.1) − 2.1 (19.9) 0.11 3.80 9.51 11.70 32.42 3.04 0.66 (0.50–0.80) Worse 15 41.8 (18.6) 37.3 (20.4) 4.4 (13.9) 0.24 3.72 9.3 8.06 22.33 3.33 0.82 (0.64–0.93) Better 21 40.3 (20.4) 39.4 (25.1) 0.9 (18.6) 0.05 4.08 10.19 10.76 29.85 3.76 0.77 (0.59–0.89) Total 75 35.3 (19.8) 35.2 (22.1) 0.1 (18.5) 0.003 3.97 9.92 10.74 29.78 2.00 0.73 (0.64–0.81) PD Psychological distress, BA Barriers to activity, DE Disinhibited eating, ES effect size, SD standard deviation, SEM standard error of measurement, SDC small detectable change, ICC intraclass correlation coefficient Page 8 of 12 Benazizi et al. Health Qual Life Outcomes (2021) 19:189 Page 9 of 12 Table 3 Convergent validity: correlation (Spearman’s r) between the Diabetes Health Profile-18 dimensions and the SF12v2 dimensions SF-12v2 dimensions Diabetes Health Profile (DHP-18) Psychological distress Barriers to activity Disinhibited eating Physical functioning − 0.268 − 0.316 − 0.322 Role-physical − 0.021 − 0.435 − 0.284 Bodily pain − 0.246 − 0.311 − 0.268 General health − 0.380 − 0.240 − 0.326 Vitality − 0.334 − 0.248 − 0.263 Social functioning − 0.248 − 0.261 − 0.236 Role-emotional − 0.353 − 0.407 − 0.412 Mental health − 0.517 − 0.292 − 0.322 All values were significant with p < 0.05; Values > 0.40 in bold Table 4 Known group validity of the DHP-18 Hypothesis Category n (%) Psychological Barriers to activity Disinhibited eating Global distress Mean (sd) p-value Mean (sd) p-value Mean (sd) p-value Mean (sd) p-value H1 Duration of disease 0.13 0.94 0.31 0.33 ≤ 5 years 54 (37.0) 31.17 (25.06) 39.74 (22.05) 37.61 (22.89) 36.29 (20.11) 6–10 years 45 (30.8) 26.42 (20.86) 38.75 (14.84) 37.19 (20.46) 34.20 (14.85) > 10 years 47 (32.2) 22.81 (14.75) 38.45 (18.29) 31.70 (19.28) 31.36 (13.45) H2 BMI 0.12 0.46 0.01 0.09 < 30 kg/m2 73 (60.8) 25.19 (18.91) 38.59 (18.78) 31.85 (20.46) 32.25 (15.92.) ≥ 30 kg/m2 47 (39.2) 29.79 (24.06) 38.94 (18.66) 40.63 (21.58) 36.36. (17.60) H3 Sex 0.08 0.04 0.06 0.03 Female 85 (58.2) 29.08 (20.89) 41.30 (18.17) 37.92 (21.34) 36.29 (15.99) Male 61 (41.8) 24.13 (21.10) 35.83 (30.92) 32.31 (20.39) 30.95 (17.13) H4 Comorbidities 0.80 0.71 0.45 0.71 No 62 (42.5) 28.76 (24.01) 40.04 (19.52) 35.34 (23.08) 34.97 (18.95) Yes 84 (57.5) 25.73 (18.61) 38.27 (18.21) 35.75 (19.58) 33.39 (14.76) H5 Educationallevel 0.85 0.75 0.89 Basic education or below 83 (56.9) 28.58 (21.97) 40.81 (17.93) 0.91 36.63 (20.98) 35.57 (31.86) Higher education 63 (43.1) 24.96 (19.75) 36.66 (19.64.) 34.20 (21.26) 32.07 (16.06) H6 Diabetes complications 0.29 0.03 0.10 0.07 No 115 (78.8) 26.52 (20.86) 37.49 (18.56) 34.44 (21.40) 32.99 (16.49) Yes 31 (21.2) 28.85 (21.96) 44.67 (18.57) 39.78 (19.49) 38.04 (16.79) recommended ranges in psychometric validation studies, was related to the emotional role dimension and not to which could be considered a strength of our study [49]. the physical role dimension as was hypothesized based Regarding convergent validity, a strong correlation was on other studies [21]. shown between the dimension of psychological distress Discriminant validity showed adequate correlations of the DHP-18 and the mental health dimension of the between the 3 dimensions, higher than those indicated in SF-12v2 and between the dimension of barriers to activ- the literature. These results differ from other studies that ity of the DHP-18 and the role physical dimension of the showed an overall low correlation between the dimen- SF-12v2, corroborating two of the predefined hypoth- sions of DHP-18 [20, 23]. eses. Similar results have been observed in previous stud- Regarding the known group validity, our results showed ies [21, 23]. However, the uninhibited eating dimension the expected trend in three of the 6 initial hypotheses. Benazizi et al. Health Qual Life Outcomes (2021) 19:189 Page 10 of 12 Regarding the hypothesis related to educational level, the possible changes produced and there may be fac- lower scores were given for the dimension barriers to tors external to the disease that can influence the results, activity with the increase in educational level. Regarding especially of repeatability-concordance, due to changes comorbidities there were also significant differences for in the context produced quickly and that can affect the specific comorbidities such as hypertension and depres- quality of life of patients. Another limitation is that sion. These results are corroborated with other stud- SF12v2 summary scores for physical and mental health ies where it was seen that the presence of hypertension can be misleading if proprietary scores are used, as a low resulted in a significantly lower score in the disinhibited physical health summary score tends to inflate the men- eating dimension [50]. In the case of the duration of the tal health summary score and vice versa. So this must disease, and despite the fact that the differences were not be taken into account when interpreting the results [53, significant, we did see that people with a disease of longer 54]. Despite this, the results are significant and similar to duration reported a better quality of life. One possible those obtained in other studies. explanation is that the longer the disease lasts, the more likely the patient is able to adapt to the care requirements including behaviour modification [51, 52]. Conclusions The CFA indicated an adequate fit to the original The strength of this study lies in the fact that this is the three-factor model, with the exception of the SRMR first adaptation and validation of a questionnaire to indicator. The dimension barriers to activity showed assess the quality of life in diabetic patients in Ecuador. the lowest standardized factorial loads, while for the Hence, it provides a practical tool to evaluate aspects dimensions psychological distress and uninhibited eat- such as self-control of food intake, limitations, barriers ing, they were adequate. Using a one-factor model, two and anxiety related to daily activities, feelings, emotions, of the 18 items, both of the activity barriers dimension, mood and irritability in people with diabetes. were loaded below the recommended value of 0.3, item The study adds to the evidence for DHP-18, showing 3 loaded with a value of 0.3 and item 1 with a value less that it is a short, acceptable, valid and reliable instru- than 0.3. Regarding item 1, the problem may be due to a ment to measure the impact of living with diabetes from lack of understanding, perhaps not at the linguistic level, a patient perspective. However, future analysis should but of concepts, since in the linguistic comprehension verify the applicability and cultural equivalence of some and cultural adaptation interviews, it was observed that of the items of barriers to activity dimension to Ecuador. the question was simple, short and easily understood, but Using DHP-18 enables clinicians to conduct an appropri- sometimes people were unsure whether food “controlling ate educational or therapeutic intervention to alleviate one’s life” referred to the need to observe and take care of or address dysfunctional life outcomes for people living one’s diet, or whether it referred to having your life struc- with diabetes. tured and organized around food, such as the timing of meals, the physical exercise depending on the amount of Abbreviations food, etc. Perhaps a clarification with examples could be PD: Psychological distress; BA: Barriers to activity; DE: Disinhibited eating; added to overcome this issue in future uses of the ques- DM: Diabetes mellitus; T2DM: Type 2 Diabetes mellitus; DHP: Diabetes Health tionnaire in Ecuador. Item 3, about “tied to meal times” Profile; DHP‐18: Diabetes Health Profile-18; CFA: Confirmatory factor analysis; DWLS: Diagonally Weighted Least Squares; TLI: Tucker-Lewis index; CFI: Com‑ was also flagged as potentially problematic in the initial parative fit index; RMSEA: Root mean square error of approximation; SRMR: round of reviews by the 2 medical researchers, who eval- Standardized root mean square residual; ICC: Intraclass correlation coefficient; uated linguistic understanding and cultural adaptation. It SEM: Standard error of measurement; SDC: Smallest detectable change; MID: Minimally important difference; SD: Standard deviation; ES: Effect size. should be added that we carried out a new confirmatory factor analysis, eliminating these 2 items and observed a Acknowledgements significant improvement in all the indicators, including The authors wish to thank María Hernández, Blanca Chamorro, Sofía Mos‑ quera, and Cristian Cuhunay for interviewing the participants. To Jimmy Martin the SRMR, which was the only one that showed a value Delgado for his help with the translation of the questionnaire. To Amparito slightly higher than recommended. Carrera, the president of the Chimbacalle diabetic patients club. To the health This study has some limitations in addition to the fac- promoters of the participating health centers. And a special thanks to all the study participants. tors discussed above. Although the DHP-18 can be used with people with either Type 1 or Type 2 diabetes, the Authors’ contributions psychometric test was not performed in type 1 diabetes, LAP and IB designed the study and all authors reviewed and approved the proposed methodology. IB, AP-C, AA-J and MCB-S carried out linguistical limiting the applicability of the results to those patients and cultural adaptation of the questionnaire. IB was responsible for data with type 2 diabetes. In addition, an important factor to acquisition in the psychometric validation. IB, MCB-S, YP and AR analysed and take into account is the context in which the study was interpreted the data. IB, MCB-S and LAP drafted the manuscript, which was critically reviewed and discussed with all co-authors. All authors read and carried out, in a pandemic situation it is difficult to assess approved the final manuscript. Benazizi et al. Health Qual Life Outcomes (2021) 19:189 Page 11 of 12 Funding 11. Saleh F, Ara F, Mumu SJ, Hafez MA. Assessment of health-related quality This research was funded by a H2020 European Research Council 2018 Start‑ of life of Bangladeshi patients with type 2 diabetes using the EQ-5D: a ing Grant, Grant Number 804761—CEAD. cross-sectional study. BMC Res Notes. 2015;8:497. 12. Zhou T, Guan H, Yao J, Xiong X, Ma A. The quality of life in Chinese popu‑ Availability of data and materials lation with chronic non-communicable diseases according to EQ-5D-3L: The datasets analysed during the current study are available from the cor‑ a systematic review. Qual Life Res Int J Qual Life Asp Treat Care Rehabil. responding author on reasonable request. 2018;27(11):2799–814. 13. Pequeno NPF, Cabral NLDA, Marchioni DM, Lima SCVC, Lyra CDO. Quality of life assessment instruments for adults: a systematic review of Declarations population-based studies. Health Qual Life Outcomes. 2020;18(1):208. 14. Ware JE, Gandek B, Guyer R, Deng N. Standardizing disease-specific qual‑ Ethics approval and consent to participate initial evaluation of the QOL Disease Impact Scale (QDIS®). Health Qual ity of life measures across multiple chronic conditions: development and All participants provided informed consent. The study protocol was approved by the ethical board at the Universidad Central del Ecuador (UCE, reference Life Outcomes. 2016;14:84. 00022-UMHE-E-2019). 15. Romero-Naranjo F, Espinosa-Uquillas C, Gordillo-Altamirano F, Barrera- Guarderas F. Which factors may reduce the health-related quality of life of Consent for publication Ecuadorian patients with diabetes? Proc R Health Sci J. 2019;38(2):102–8. All authors fulfil the criteria for authorship and have read and approved the 16. Lara M, José M. Diabetes mellitus y sus factores de riesgo en el Ecuador. final version of this manuscript. agosto de 2016 [cited 2021 Jan 21]. http://​repos​itorio.​usfq.​edu.​ec/​han‑ dle/​23000/​5697 Competing interests 17. Jácome Á, Francisco J. Factores de riesgo socioeconómicos en la The authors declare that they have no competing interests. prevalencia de diabetes tipo II: evidencia en el Ecuador ENSANUT-ECU 2011–2013. 2018 [cited 2021 Jan 30]. http://​repos​itorio.​puce.​edu.​ec:​80/​ Author details xmlui/​handle/​22000/​15622 1 Department of Public Health, Universidad Miguel Hernández, Sant Joan 18. Pereira EV, Tonin FS, Carneiro J, Pontarolo R, Wiens A. Evaluation of the d’Alacant, Alicante, Spain. 2 CIBER de Epidemiología y Salud Pública (CIBER‑ application of the Diabetes Quality of Life Questionnaire in patients with ESP), Madrid, Spain. 3 Health Services Research Group, IMIM (Hospital del diabetes mellitus. Arch Endocrinol Metab. 2020;64(1):59–65. Mar Medical Research Institute), Barcelona, Spain. 4 Universitat Autònoma de 19. Palamenghi L, Carlucci MM, Graffigna G. Measuring the quality of life in Barcelona, Barcelona, Spain. 5 Cardiovascular Epidemiology and Research Unit, diabetic patients: a scoping review. J Diabetes Res. 2020;2020:5419298. University Hospital and Research Institute Vall d’Hebron (VHIR), Barcelona, 20. Meadows K, Steen N, McColl E, Eccles M, Shiels C, Hewison J, et al. The Spain. 6 Instituto de Salud Pública, Pontificia Universidad Católica del Ecuador, Diabetes Health Profile (DHP): a new instrument for assessing the psycho‑ Quito, Ecuador. 7 Unidad Docente de Medicina Preventiva y Salud Pública de social profile of insulin requiring patients–development and psycho‑ Cantabria, Consejería de Sanidad de Cantabria, Santander, Spain. 8 Department metric evaluation. Qual Life Res Int J Qual Life Asp Treat Care Rehabil. of Experimental and Health Sciences, Pompeu Fabra University, Barcelona, 1996;5(2):242–54. Spain. 21. Santos Cruz R, Leitão CE, Lopes FP. Determinantes do estado de saúde dos diabéticos. Rev Port Endocrinol Diabetes E Metab. 2016;11(2):188–96. Received: 12 April 2021 Accepted: 6 July 2021 22. Tan ML, Khoo EY, Griva K, Lee YS, Amir M, Zuniga YL, et al. Diabetes health profile-18 is reliable, valid and sensitive in Singapore. Ann Acad Med. 2016;45(9):383–93. 23. Jelsness-Jørgensen L-P, Jensen Ø, Gibbs C, Bekkhus Moe R, Hofsø D, Bernklev T. Psychometric testing of the Norwegian Diabetes Health References Profile (DHP-18) in patients with type 1 diabetes. BMJ Open Diabetes Res 1. World Health Organization. Diabetes. [cited 2021 Jan 21]. https://​www.​ Care. 2018;6(1):e000541. who.​int/​es/​news-​room/​fact-​sheets/​detail/​diabe​tes 24. García-Carrera C, Gutierrez-Fuentes E, Borroel-Saligan L, Oramas P, Vidal- 2. Orces CH, Lorenzo C. Prevalence of prediabetes and diabetes among López M. Club de diabéticos y su impacto en la disminución de glicemia older adults in Ecuador: analysis of the SABE survey. Diabetes Metab del diabético tipo 2. Salud en Tabasco. 2002;8(1):16–9. Syndr. 2018;12(2):147–53. 25. Olvera JP. La influencia del grupo de autoayuda de pacientes diabéticos 3. Cordero LCA, C MAV, Cordero G, Álvarez R, Añez R, Rojas J, et al. Prevalen‑ en el control de su enfermedad. Horiz Sanit. 2009;8(1):44–58. cia de la diabetes mellitus tipo 2 y sus factores de riesgo en individuos 26. Gandek B, Ware JE, Aaronson NK, Apolone G, Bjorner JB, Brazier JE, et al. adultos de la ciudad de Cuenca-Ecuador. Av En Biomed. 2017;6(1):10–21. Cross-validation of item selection and scoring for the SF-12 Health Survey 4. Pan American Health Organization / World Health Organization. La diabe‑ in nine countries: results from the IQOLA Project. International Quality of tes, un problema prioritario de salud pública en el Ecuador y la región de Life Assessment. J Clin Epidemiol. 1998;51(11):1171–8. las Américas. 2014 [cited 2021 Jan 21]. https://​www.​paho.​org/​ecu/​index.​ 27. Omary MB, Eswaraka J, Kimball SD, Moghe PV, Panettieri RA, Scotto php?​option=​com_​conte​nt&​view=​artic​le&​id=​1400:​la-​diabe​tes-​un-​probl​ KW. The COVID-19 pandemic and research shutdown: staying safe and ema-​prior​itario-​de-​salud-​publi​ca-​en-​el-​ecuad​or-y-​la-​regio​nde-​las-​ameri​ productive. J Clin Investig. 2020;130(6):2745–8. cas&​Itemid=​360 28. Mokkink LB, Prinsen CAC, Bouter LM, de Vet HCW, Terwee CB. The 5. Organización Mundial de la Salud – Perfiles de los países para la diabetes. COnsensus-based Standards for the selection of health Measurement 2016 [cited 2021 Jan 21]. https://​www.​who.​int/​diabe​tes/​count​r ypro​files/​ INstruments (COSMIN) and how to select an outcome measurement ecu_​es.​pdf?​ua=1 instrument. Braz J Phys Ther. 2016;20(2):105–13. 6. MSP, INEC, OPS/OMS. ENCUESTA STEPS ECUADOR. 2018 [cited 2021 Jan 29. Bell ML, Fairclough DL, Fiero MH, Butow PN. Handling missing items in 21]. https://​www.​salud.​gob.​ec/​wpcon​tent/​uploa​ds/​2020/​10/​INFOR​ the Hospital Anxiety and Depression Scale (HADS): a simulation study. MESTE​PS.​pdf BMC Res Notes. 2016;9(1):479. 7. Glovaci D, Fan W, Wong ND. Epidemiology of diabetes mellitus and 30. Meadows K. Diabetes health profile user manual—sample pages. 2015. cardiovascular disease. Curr Cardiol Rep. 2019;21(4):21. 31. McHorney CA, Tarlov AR. Individual-patient monitoring in clinical prac‑ 8. Solli O, Stavem K, Kristiansen IS. Health-related quality of life in diabetes: tice: are available health status surveys adequate? Qual Life Res Int J Qual the associations of complications with EQ-5D scores. Health Qual Life Life Asp Treat Care Rehabil. 1995;4(4):293–307. Outcomes. 2010;8:18. 32. Meadows KA, Abrams C, Sandbaek A. Adaptation of the Diabetes Health 9. Rubin RR, Peyrot M. Quality of life and diabetes. Diabetes Metab Res Rev. Profile (DHP-1) for use with patients with Type 2 diabetes mellitus: 1999;15(3):205–18. psychometric evaluation and cross-cultural comparison. Diabet Med J Br 10. Cannon A, Handelsman Y, Heile M, Shannon M. Burden of illness in type 2 Diabet Assoc. 2000;17(8):572–80. diabetes mellitus. J Manag Care Spec Pharm. 2018;24(9-a Suppl):S5-13. Benazizi et al. Health Qual Life Outcomes (2021) 19:189 Page 12 of 12 33. Mîndrilă D. Maximum likelihood (ML) and diagonally weighted least loss behaviors in the United States. Diabetes Metab Syndr Obes Targets squares (DWLS) estimation procedures: a comparison of estimation Ther. 2020;13:2049–55. bias with ordinal and multivariate non-normal data. Int J Digit Soc. 46. Alshayban D, Joseph R. Health-related quality of life among patients 2010;1(1):60–6. with type 2 diabetes mellitus in Eastern Province, Saudi Arabia: a cross- 34. Flora DB, Curran PJ. An empirical evaluation of alternative methods of sectional study. PLoS ONE. 2020;15(1):e0227573. estimation for confirmatory factor analysis with ordinal data. Psychol 47. Norris SL. Health-related quality of life among adults with diabetes. Curr Methods. 2004;9(4):466–91. Diab Rep. 2005;5(2):124–30. 35. Forero CG, Maydeu-Olivares A, Gallardo-Pujol D. Factor analysis with ordi‑ 48. Glasgow RE, Ruggiero L, Eakin EG, Dryfoos J, Chobanian L. Quality of life nal indicators: a Monte Carlo study comparing DWLS and ULS estimation. and associated characteristics in a large national sample of adults with Struct Equ Model Multidiscip J. 2009;16(4):625–41. diabetes. Diabetes Care. 1997;20(4):562–7. 36. Prinsen CAC, Vohra S, Rose MR, Boers M, Tugwell P, Clarke M, et al. How to 49. Polit DF. Getting serious about test–retest reliability: a critique of retest select outcome measurement instruments for outcomes included in a research and some recommendations. Qual Life Res. 2014;23(6):1713–20. “Core Outcome Set”—a practical guideline. Trials. 2016;17(1):449. 50. Goddijn P, Bilo H, Meadows K, Groenier K, Feskens E, Meyboom-de JB. 37. Hu L, Bentler PM. Cutoff criteria for fit indexes in covariance structure The validity and reliability of the Diabetes Health Profile (DHP) in NIDDM analysis: conventional criteria versus new alternatives. Struct Equ Model. patients referred for insulin therapy. Qual Life Res Int J Qual Life Asp Treat 1999;6(1):1–55. Care Rehabil. 1996;5(4):433–42. 38. Clark LA, Watson D. Constructing validity: basic issues in objective scale 51. Lambrinou E, Hansen TB, Beulens JW. Lifestyle factors, self-manage‑ development. Psychol Assess. 1995;7(3):309–19. ment and patient empowerment in diabetes care. Eur J Prev Cardiol. 39. Mulhern B, Meadows K. Investigating the minimally important difference 2019;26(2_suppl):55–63. of the Diabetes Health Profile (DHP-18) and the EQ-5D and SF-6D in a UK 52. Chrvala CA, Sherr D, Lipman RD. Diabetes self-management education for diabetes mellitus population. Health (NY). 2013;05(06):1045–54. adults with type 2 diabetes mellitus: a systematic review of the effect on 40. Terwee CB, Bot SDM, de Boer MR, van der Windt DAWM, Knol DL, Dekker glycemic control. Patient Educ Couns. 2016;99(6):926–43. J, et al. Quality criteria were proposed for measurement properties of 53. Farivar SS, Cunningham WE, Hays RD. Correlated physical and mental health status questionnaires. J Clin Epidemiol. 2007;60(1):34–42. health summary scores for the SF-36 and SF-12 Health Survey, V.I. Health 41. Bland JM, Altman DG. Measurement error. BMJ. 1996;312(7047):1654. Qual Life Outcomes. 2007;5:54. 42. Kazis LE, Anderson JJ, Meenan RF. Effect sizes for interpreting changes in 54. Fleishman JA, Selim AJ, Kazis LE. Deriving SF-12v2 physical and mental health status. Med Care. 1989;27(3 Suppl):S178-189. health summary scores: a comparison of different scoring algorithms. 43. Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Qual Life Res Int J Qual Life Asp Treat Care Rehabil. 2010;19(2):231–41. Hillsdale: L. Erlbaum Associates; 1988. 44. Jing X, Chen J, Dong Y, Han D, Zhao H, Wang X, et al. Related factors of quality of life of type 2 diabetes patients: a systematic review and meta- Publisher’s Note analysis. Health Qual Life Outcomes. 2018;16(1):189. Springer Nature remains neutral with regard to jurisdictional claims in pub‑ 45. Rozjabek H, Fastenau J, LaPrade A, Sternbach N. Adult obesity and health- lished maps and institutional affiliations. related quality of life, patient activation, work productivity, and weight Ready to submit your research ? Choose BMC and benefit from: • fast, convenient online submission • thorough peer review by experienced researchers in your field • rapid publication on acceptance • support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations • maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions