Translate this page into:
Spoken knowledge in low literacy diabetes scale: Reliability and validity assessment on Indian Type 2 diabetes patients
-
Received: ,
Accepted: ,
How to cite this article: Phukan P. Spoken knowledge in low literacy diabetes scale: Reliability and validity assessment on Indian Type 2 diabetes patients. Indian J Med Sci 2022;74:152-4.
Abstract
India is facing a diabetes epidemic and it urgently needs to address the health literacy needs of diabetes patients to reduce the catastrophic impact of the disease. Since a valid and reliable tool for assessing diabetes literacy assessment was lacking, this study was carried out to conduct a psychometric analysis of the Spoken Knowledge in Low Literacy Diabetes (SKILLD) Scale among Indian Type 2 diabetes patients. The study participants belonged mostly to the lower socioeconomic strata with low literacy. The results of the Kannada version of the SKILLD scale administered to 100 Type-2 diabetes patients visiting a tertiary care center in South India are presented here.
Keywords
Low literacy
Type-2 diabetes
Psychometric analysis
India
INTRODUCTION
A dual problem exists in diabetes management, which is a lack of diabetes literacy among patients[1,2] and the inability of patients to understand physicians’ instructions.[3,4] This results in poor diabetes outcomes. Diabetes literacy assessments are now mandatory at the initial stages of patient evaluation as a way of ensuring the quality of diabetes care.[5] However, the treating physician will need a tool that can be used to assess patients’ knowledge and ability to follow diabetes related instructions.
THE STUDY INSTRUMENT
The Spoken Knowledge in Low Literacy Diabetes (SKILLD) was developed by Rothman et al.[6] to aid in the assessment of diabetes literacy among low literacy patients. The 10-item SKILLD assesses core knowledge related to self-care, including glucose management, appropriate lifestyle modifications, the recognition and treatment of acute complications, and appropriate activities to prevent long-term consequences of uncontrolled disease. Each item carries 10 points. The total score ranges from 0% to 100%. A total SKILLD score of >50% indicated good diabetes literacy. The expected response for each item was derived from the new American Diabetes Association (2020) standards for diabetes care[5] [Table 1].
| Questions (Score-0 or 10 for each) | Good knowledge *(n=100) | Median SKILLD Score | P-value¥ |
|---|---|---|---|
| What are the signs and symptoms of high blood sugar? How do you feel when your blood sugar is high or when you were diagnosed? | 77 | 10 | 0.002 |
| What are the signs and symptoms of low blood sugar? How do you feel when your blood sugar is too low? | 84 | 10 | 0.14 |
| How do you treat low blood sugar? What should you do if your sugar is too low? How can you bring your blood sugar up if it is too low? | 42 | 5 | 0.21 |
| How often should a person with diabetes check his or her feet? Once a day, once a week, or once a month? | 63 | 5 | 0.48 |
| Why are foot exams important in someone with diabetes? Why is it important to look at your feet? What are you looking for? | 47 | 5 | 0.01 |
| How often should you see an eye doctor and why is it important? How often? Why? | 34 | 0 | 0.79 |
| What is a normal fasting blood glucose or blood sugar? When you get up first thing in the morning and check your blood sugar before you eat or take medicine, what should it be? What 2 numbers? | 57 | 5 | 0.002 |
| What is a normal HbA1c (hemoglobin A1C) or “average blood sugar test”? When they draw blood from your arm and get an average blood sugar reading, what should it be? | 13 | 0 | 0.32 |
| How many times per week should someone with diabetes exercise and for how long? How many times a week? How long or how much per day? | 63 | 5 | 0.61 |
| What are some long-term complications of uncontrolled diabetes? Do you know anyone that has diabetes and had “bad things” happen to them? What are some of those “bad things”? | 81 | 10 | 0.14 |
| Overall Good Knowledge | 50 | 50 | 0.03 |
This study was undertaken to conduct a psychometric assessment of the SKILLD Scale among Kannada speaking Type-2 diabetes patients in South India and to ascertain the socio-clinical correlates of diabetes. Inclusion criteria were Type-2 diabetes patients with a history of diabetes for at least 6 months and age 20 years or older.
PROCESS OF PSYCHOMETRIC ANALYSIS
The English version was translated into the Kannada version by back translation procedures by two social scientists as per the Brislin translation model[7] after obtaining necessary permission from the authors. Knowledge scores were compared with the patient’s education level and a clinical parameter related to glycemic control. Those with HbA1c ≤7% are considered to have good glycemic control in this study.
RELIABILITY AND VALIDITY RESULTS
Kuder-Richardson coefficient was 0.73 (0.66–0.81, 95% CI), indicating good reliability of results. The item total correlation ranged from 0.25 to 0.56. The mean total SKILLD score was 44.6% (SD=14.71, SEM=1.47, Range−0–65%, 95% CI).
Factor analysis
Kaiser-Meyer-Olkin measure was 0.76 and Bartlett’s test showed the adequacy of sample size (χ2 = 197.752, df = 45, P < 0.0001), suggesting that the dataset was fit for factor analysis. Principal component analysis with Promax rotation revealed three-factor structure with Eigenvalue >1 and explained 31.7%, 46.8%, and 57.4% total variance, respectively. The theoretical constructs of the three factors are identified as Factor 1: Knowledge related to treatment and self-care. (Items − 3, 4, 6, 9, and 10), Factor 2: Knowledge related to diagnosis of diabetes related signs and symptoms (Items − 1, 2, and 5), Factor 3: Knowledge related to diabetes numeracy for blood sugar monitoring, that is, the normal value of fasting blood sugar and HbA1c (Items 7 and 8). However, the subscale reliability coefficient for these three factors was low at 0.65, 0.67, and 0.45, respectively. The original study had reported only one factor structure.[6]
Construct validity
The mean Total SKILLD score was higher (54.95%) among the higher education category compared to those with less than 6th Standard education (41.87%) and it was statistically significant at P = 0.03 as shown in [Table 1]. Further, a significant correlation was found between total SKILLD scores with education (r = 0.41, 0.2–0.6, 95% CI, P < 0.001) and income categories (r = 0.25, 0.01−0.47, 95% CI, P = 0.02). ANOVA revealed significant difference in education level and SKILLD scores (0 = No knowledge, 5 = partial knowledge, and 10=good knowledge), (F(1,98) = 7.20, P = 0.009). The eta squared was 0.07 indicating the education levels explain approximately 7% of the variance in SKILLD scores. Post hoc paired t-test Tukey pair-wise comparison showed that mean SKILLD scores were higher (>50%) among higher education category (Mean=47.35, SD=12.47, n = 66) was significantly larger than for lower education category (M=39.26, SD=17.28, n = 34) at P = 0.009.
[Table 2] shows that there was no significant difference in knowledge level when compared with HbA1c (χ2(1) = 0.21, P = 0.64) and income (χ2(1) = 0.21, P = 0.64). However, there was significant difference in diabetes knowledge level and was found in relation to duration of diabetes (χ2(1) = 8.02, P < 0.005), gender (χ2(1) = 7.41, P < 0.006), level of education (χ2(1) = 18.75, P < 0.001), BMI (χ2(1) = 77, P = 0.005), HDL: Cholesterol ratio (χ2(1) = 9.80, P = 0.02), and type of medication (χ2(1) = 32.67, P < 0.001).
| Variable | SKILLD score | *P-value | |
|---|---|---|---|
| Low knowledge ≤50 | Good knowledge >50 | ||
| HbA1c | |||
| ≤ 7% | 20 | 12 | 0.64 |
| > 7% | 40 | 23 | |
| Duration of diabetes | |||
| < 5 years | 26 | 17 | 0.005 |
| > 5 years | 38 | 19 | |
| Gender | |||
| Female | 37 | 19 | 0.006 |
| Male | 27 | 17 | |
| Education | |||
| Up to 6thStandard | 25 | 9 | <0.001 |
| Higher | 39 | 27 | |
| Income | |||
| < Rs 10,000 | 44 | 23 | 0.64 |
| > Rs10,000 | 20 | 13 | |
| BMI | |||
| Normal | 12 | 6 | 0.005 |
| High | 52 | 30 | |
| HDL: Cholesterol Ratio | |||
| Normal | 8 | 3 | 0.02 |
| High | 17 | 11 | |
| Type of Medication | |||
| Oral | 48 | 30 | <0.001 |
| Both Oral and Insulin | 16 | 6 | |
Concurrent validity
There was a low negative correlation between HbA1c level and total skilled score (r = 0.01, −0.21 and 0.19, 95% CI) but not statistically significant (P = 0.94) which might be because the majority of the patients (63%) in this study had poor glycemic control (HbA1c >7%). Chi-square goodness of fit test revealed that patients with low SKILLD scores had significantly higher HbA1c than expected [χ2(1) = 6.67, P < 0.01].
PATIENT EDUCATION AND KNOWLEDGE EVALUATION
Low literacy is still widely prevalent in India,[8,9] which means that there are substantial numbers of people who might not be able to understand physicians’ instructions and contributing to poor diabetes outcomes. Studies from India also concur that the diabetes-related morbidities and mortalities are attributed to poor diabetes literacy.[9,10] This study results also confirm that the majority of the patients did not have information on diabetes self-care. Item 7 and Item 8 on SKILLD scale captured numeracy skills deficit related to normal fasting blood glucose and HbA1c levels which are critical in diabetes self-care practices.[5,6] Researchers can use this instrument to gather more data on numerical skills among patients. However, the SKILLD scale does not have items related to knowledge of nutrition and diabetes medication. It does measure most of the skills necessary for daily self-care. The tool will help both physicians and patients to empower themselves with the necessary health information needed to improve the quality of diabetes care.
The tool could clearly define the areas of knowledge deficit and help in patient education interventions. The SKILLD tool can be used both as a diagnostic tool as well as an evaluation tool in diabetes health literacy. India is a country with linguistic and cultural diversity. Therefore, psychometric analysis of this instrument in other Indian languages and in various settings and on a larger population is required to gather more evidence.
Limitations
This study being a cross-sectional study test-retest reliability analysis could not be done.
Acknowledgment
I thank the patients for their participation and for the cooperation and support of the institution in conducting the study.
Declaration of patient consent
Institutional Review Board (IRB) permission was obtained for the study.
Conflicts of interest
There are no conflicts of interest.
Financial support and sponsorship
Nil.
References
- Health literacy, diabetes prevention and self-management. J Diabetes Res. 2017;2017:1298315.
- [CrossRef] [PubMed] [Google Scholar]
- Health literacy status and understanding of the prescription instructions in diabetic patients. J Diabetes Res 2018:1-5.
- [CrossRef] [PubMed] [Google Scholar]
- Health literacy in diabetes care: Explanation, evidence, and equipment. Diabetes Manag (Lond). 2011;1:191-9.
- [CrossRef] [PubMed] [Google Scholar]
- What did the doctor say? Health literacy and recall of medical instructions. Med Care. 2012;50:277-82.
- [CrossRef] [PubMed] [Google Scholar]
- The spoken knowledge in low literacy in diabetes scale: A diabetes knowledge scale for vulnerable patients. Diabetes Educ. 2005;31:215-24.
- [CrossRef] [PubMed] [Google Scholar]
- Back-translation for cross-cultural research. J Cross Cult Psychol. 1970;1:187-216.
- [CrossRef] [Google Scholar]
- The increasing burden of diabetes and variations among the states of India: The Global Burden of Disease Study 1990-2016. Lancet Glob Health. 2018;6:e1352-62.
- [CrossRef] [Google Scholar]
- Current status of management, control, complications and psychosocial aspects of patients with diabetes in India: Results from the DiabCare India 2011 Study. Indian J Endocr Metab. 2014;18:370-8.
- [CrossRef] [PubMed] [Google Scholar]
- Level of health literacy among Type 2 diabetic persons and its relation to glycemic control. J Assoc Physicians India. 2019;67:59-62.
- [Google Scholar]
