Insulin usage technique and associated factors: a cross-sectional study in Southern Brazil

¹Health Program, Life Cycle, Health Sciences Center, Catholic University of Pelotas, Pelotas, Brazil, ²Epidemiology Program, Faculty of Medicine, Federal University of Pelotas, Pelotas, Brazil, ³Faculty of Medicine, Health Sciences Center, Catholic University of Pelotas, Pelotas, Brazil

^&Corresponding author

Introduction    

Diabetes mellitus is a global epidemic posing major challenges to health systems worldwide [1]. An estimated 589 million adults worldwide aged 20 to 79 live with diabetes, and this number is expected to increase to 853 million by 2050 [2]. The factors responsible for the increase include an aging population, urbanization, and the growing prevalence of obesity and sedentary lifestyles [3]. Brazil has one of the highest prevalence rates in the world, the highest in Latin America, and ranks sixth globally in number of cases. From 2013 to 2023, 64,710 adults in Brazil died from diabetes complications, with fatality rates ranging from 9.7% in 2013 to 8.2% in 2023 [4].

Insulin therapy aims to mimic the physiological profile of pancreatic insulin secretion, providing glycemic control, which is essential for preventing acute and chronic complications of the disease [5]. Although insulin is highly effective, it must be correctly used to achieve its therapeutic effects. This involves patients storing, preparing, and administering the dose in a home environment, which is a crucial aspect of disease control [6]. However, there is evidence that 75% of patients with diabetes do not administer and/or store disposable insulin pens correctly [7]. In Brazil, a study published in 2009 showed that only 61% of patients properly prepare and administer insulin [8]. Despite the growing prevalence of the disease and the fact that in Brazil, the Unified Health System (SUS - Sistema Unico de Saude in Portuguese) provides NPH and regular human insulin in 10 mL vials and pens for free to all patients with diabetes [9], no recent studies planned to assess insulin use technique was found in the literature. Thus, the objective of this study was to assess the storage, preparation, and application of insulin among diabetic patients living in the coverage areas of the Primary Health Care System in Canguçu, a city in South Brazil.

Methods     

Study design: this was a cross-sectional study, carried out from December 2024 to January 2025.

Setting: recognized as the national capital of family farming, with a large number of smallholdings, Canguçu has a population of 49,680 inhabitants [10]. Situated in the southern region, 274 km from the Rio Grande do Sul state capital city (Porto Alegre), it has a strong influence from Azorean, German, and Italian colonization.

Participants: eligible individuals were adults aged 18 years or older with type 1 or 2 diabetes who lived in the urban area of Canguçu and got insulin from the Municipal pharmacy in the previous year. The insulin delivered by the municipal pharmacy is provided for free by the Brazilian Unified Health System to individuals living with diabetes. The home addresses of the individuals on the list were subsequently provided to the research team by the Municipal Health Department.

Variables: the outcome variable was the correct technique for insulin use, including storage, preparation, and application, as defined by the Ministry of Health [11] and the Brazilian Diabetes Society [12] (Annex 1).

Data sources/measurement: to identify the adequacy of storage, the participants were asked where they usually stored the insulin they were using, and the insulin that was not yet open. To verify the adequacy of insulin preparation and administration, the participants were questioned regarding how they usually prepared and administered their insulin. The interviewee's spontaneous responses were recorded in a checklist based on the Ministry of Health and Brazilian Diabetes Society as “yes, spontaneous”, followed by the question, “anything else?” Once the spontaneous responses had been exhausted, leading questions were asked about items not mentioned. For example, if the respondents did not spontaneously mention washing their hands, they were asked explicitly about that, and the answer “no” or “yes, prompt” was recorded accordingly. Each of the items related to adequacy of storage, preparation, and application was coded as 0 (not reported spontaneously or after the leading question) or 1 (yes). In a subsample, the preparation and administration of insulin were observed during a second scheduled home visit.

Independent variables: the independent variables included the patient sociodemographic and behavioral characteristics, characteristics of the disease, place where they usually received health care, and comorbidity. Sociodemographic characteristics: family income in Brazilian reais in the month prior to the interview; self-reported ethnicity (white, brown, black, Asian, or indigenous); education (number of years of schooling completed with passing grades); household crowding (more than two residents per bedroom); age (in full years); and sex (male or female). Behavioral characteristics: smoking status (never, current, or former); alcohol consumption in the last month (yes or no); physical activity in the last month (yes or no); and weekly frequency of physical activity (days per week). Disease characteristics: age at diabetes diagnosis; disease duration in years; type of diabetes (type 1, type 2, or unknown); duration of insulin use (in years); type of insulin used (NPH, regular, or both); self-administration of insulin (yes or no); and site of insulin injection (abdomen, thighs, buttocks, or back of the arms). Place where patients usually received health care: "basic health unit (UBS) in the neighborhood," "UBS in the neighborhood + other location," or "other location"; whether they received guidance on insulin use, and who provided the guidance (physician, nurse, or other). Comorbidity: self-reported medical diagnosis of arterial hypertension, heart disease, kidney disease, or visual impairment due to diabetes.

Data collection (participant interviews) took place between December 2024 and January 2025 and was carried out at the participants' homes using a structured questionnaire with closed questions that had been previously tested, administered by trained interviewers. A thermometer was used to check the temperature of the environment in which the insulin vial or pen was located. The subsample was observed using a checklist constructed by the authors.

Bias: information bias from the fact that information on insulin storage, preparation, and administration was obtained through patient self-reporting. To overcome this limitation, during the visit the participants were asked to show where they stored their insulin. Additionally, a subsample study was conducted to directly observe how patients implemented insulin preparation and administration.

Study size: assuming an outcome prevalence of 10-50%, significance level of 95%, and a design effect equal to one (assuming the study population was homogeneous, given that all were users of the municipal pharmacy), the available sample size would allow identification of prevalence ratios (PR) of at least 3.5 for most of the exposures. The subsample of 10% of participants was selected by stratified sampling according to area of residence.

Statistical methods: the data were analyzed using the Stata 17.0 statistical program [13]. First, the independent variables were described as means, standard deviations (SD) or proportions. Next, a score was calculated for each phase of insulin use (storage, preparation, and application). Given that there was a different number of steps for preparing and administering insulin according to the type of insulin presentation (vials or pens), the scores were normalized to fit within a range of 0-100. A score was calculated for each phase of insulin use (storage, preparation, and application) by dividing the total affirmative responses by the number of items in each phase and multiplying the result by 100. Storage for both insulin presentations (vials or pens) accounted for a total of four items. Preparing insulin in vials comprised 10 items, while preparing insulin in pens comprised five items. Insulin application from syringes accounted for seven items, while application from pens accounted for nine items. Overall scores for insulin vial users were obtained by dividing the sum of the storage, preparation, and application items by 21 and multiplying the result by 100. For insulin pen users, the sum of all items was divided by 18 and multiplied by 100. The one participant who used both vial and pen insulin was excluded from these analyses. Since the distribution of partial and total scores was not normal, medians and interquartile ranges (IQRs) were calculated for each independent variable. For the same reason, the overall score was dichotomized further using the median as the cutoff point (yes/no), and those whose overall score was above the median were recorded as presenting the outcome (adequate technique of insulin use). Nonparametric Mann-Whitney and Kruskal-Wallis tests were used to compare the medians of dichotomous and categorical variables with more than two categories, respectively. Crude and adjusted PR, with 95% confidence intervals (95%CIs), were calculated using Poisson regression with robust variance. For the adjusted analysis, all independent variables were initially included, then removed one by one from highest to lowest p-value, until only those associated with the outcome at p≤ 0.20 remained in the model.

Ethical considerations: this study was approved by the Municipal Health Department of Canguçu and the Research Ethics Committee of the Catholic University of Pelotas under office number 7223403 and certificate of ethical review number 84554524.7.0000.5339. All participants signed an informed consent form before being included in the study.

Results     

Participants: of the 150 registries provided by the municipal pharmacy, 14 had died, and 11 could not be located at the address provided by the Municipal Health Department. Three of the remaining 125 refused to participate in the study, and four reported that they were not using insulin. After the exclusion of the deceased and of those who denied using insulin, the attrition rate was 10.6%, remaining 118 participants for inclusion in the analysis.

Descriptive data: Annex 1 shows the distribution of participants according to sociodemographic and behavioral characteristics. The mean age was 63.6 years (SD=13.9). The majority was female (68.6%) and self-identified as white (87.2%). More than one-fifth (23.7%) did not attend school or did not complete the first grade of elementary school. About half (48.3%) had four to eight years of schooling, and 9.3% had nine or more years of formal education. Just over 4% reported domestic crowding. Regarding behavioral characteristics, 11.9% of the participants were current smokers. Alcohol consumption in the last month was reported by 15.3% of the respondents. More than one-third (39.0%) reported engaging in physical activity, 43.5% of whom did so seven days a week.

Seven participants (5.9%) were diagnosed in childhood, and three in adolescence. Most were diagnosed at age 40 or older (Annex 1). The mean disease duration was 20.5 years (SD=12.4), and 52.5% had known they had diabetes for ≥20 years. Type 2 was reported by 42.4% of respondents, and a similar proportion were unaware of their type of diabetes. Medical diagnosis of arterial hypertension, heart disease, kidney disease, and visual impairment due to diabetes were reported, respectively, by 80.5%, 44.4%, 28.0%, and 51.7% of the participants.

Over 40% of the participants had used insulin for less than five years. Approximately three-quarters (73.9%) used NPH insulin exclusively, while 26.1% used a combination of NPH and regular insulin. The abdomen was the most frequently reported site of insulin administration. The UBS in the neighborhood was the place of consultation for diabetes management for 30.5% of the sample. The vast majority (86.4%) responded that they had received guidance on how to use insulin, and most reported that a physician had provided the guidance (71.6%).

Outcome data

Insulin storage: three participants did not have any insulin at home at the time of the interview. Of the remaining 115 participants, 85 used only NPH insulin, and 30 used a combination of regular and NPH insulin. Most participants (60.9%) used the pen as their insulin delivery method. Most participants (88.0%) usually stored their insulin in the refrigerator. The ambient temperature recorded on the thermometer placed next to the insulin outside the refrigerator averaged 26.5°C (SD=1.9°C), ranging from 23.5°C to 29.0°C. All participants reported keeping unopened insulin in the refrigerator. Only one participant recorded on the packaging the date the insulin had been opened. The insulin in use by 113 participants who used NPH insulin and by 29 participants who used regular insulin was within the expiration date stated by the manufacturer on the packaging.

Preparation and application of insulin: in the analysis of insulin preparation and administration, one participant was excluded because he used both insulin presentations, whereas our study was restricted to the use of either vials or pens exclusively. For the preparation phase, 114 participants remained: 44 used insulin vials and 70 used insulin pens. Regarding administration, data were obtained for 69 of the 70 participants who used pens, resulting in a final total of 113 participants.

The homogenization of NPH insulin in vials was spontaneously recalled by 19 participants (43.2%) (Table 1). Only two participants reported rolling the vial 20 times (both responses were spontaneous). Less than 10% spontaneously recalled washing their hands, cleaning the rubber on the vial, drawing air into the syringe, and injecting air into the vial. No interviewee spontaneously mentioned protecting the needle until insulin administration. After the leading question, less than one in five reported doing so. Except for injecting air into the vial, the frequency of all procedures increased after the leading question as expected, with hand washing (88.6%) and checking the appearance of the insulin (61.4%) being the most frequently mentioned procedures.

Table 2 shows the steps used in insulin application for the 44 participants who used insulin in vials. The 90- or 45-degree angle was the most spontaneously mentioned (N=35; 81.4%), while rotation of the application site (N=2) and keeping the needle in place for at least five seconds after insulin administration (N=2) were the least spontaneously reported. After being prompted, more than 90% confirmed that they rotated the vial and 35.6% confirmed that they kept the needle in place for at least five seconds after insulin administration. Most did not use alcohol (53.5%) or pinch the skin (55.8%) for insulin application. Seventy percent of participants reused the needle, with more than 90% doing so a maximum of eight times. Among the 70 participants (61.4%) who used insulin pens, the procedures most spontaneously remembered by this group were movements of the pen to homogenize NPH insulin and inserting the prescribed units (67.1% and 52.9%, respectively) (Table 1). Only four participants mentioned performing ten movements to homogenize the insulin. The test to check needle permeability was the least mentioned (2.9%). Handwashing (84.3%) and screwing on the needle (67.1%) were confirmed most frequently after the leading question.

As with participants who used insulin in vials, the angle of application was most frequently mentioned spontaneously (Table 2). After the leading question, the majority (85.5%) reported rotating the application site and 59.4% confirmed waiting at least ten seconds before removing the needle after insulin administration. More than three-quarters of the respondents (77.8%) reported slowly removing the needle after insulin administration. Most participants did not use alcohol (69.6%) or pinch the skin (60.9%) when applying insulin. Even after the leading question, more than half (53.6%) did not mention applying light local pressure in case of bleeding. Overall, 55 participants (80.0%) responded that they reused needles, with 44 doing so a maximum of eight times.

Main results

Insulin use adequacy scores: the median overall adequacy score for those using insulin in vials was 66.7% (IQR: 52.4-71.4) and for those using insulin in pens, 66.7% (IQR: 55.6-72.2) (Mann-Whitney p = 0.590) (Annex 1). The median overall adequacy score for the entire sample, regardless of insulin presentation type, was 66.7% (IQR: 52.4-71.4%). The prevalence of patients with an overall score above the sample median was 28.8% (95%CI 20.8%-37.9%). Table 3 shows the crude and adjusted PR for variables that remained in the final adjusted model. In the adjusted analysis, the prevalence of proper insulin use was 2.5 times higher among participants of brown, black, Asian, or indigenous ethnicity than among those who self-identified as white (PR=2.53; 95%CI 1.37-4.67; p=0.001). The prevalence increased with higher levels of education (linear trend p<0.001): it was 2.64 times higher among those with 1-4 years of schooling (95%CI 1.17-5.94) and 5.29 times higher among those with ≥5 years of schooling (95%CI 2.39-11.67) compared to those who did not receive a formal education. The PR was more than three times higher among those who reported consuming alcoholic beverages in the month prior to the interview (PR=3.31; 95%CI 1.90-5.75).

The older the age at diagnosis, the greater the likelihood of an appropriate technique. This likelihood was four times higher among those diagnosed between the ages of 20 and 49 than among those diagnosed in childhood or adolescence (linear trend p = 0.023). Among those diagnosed at age 50 or older, the PR was 9.55 (95%CI: 0.99-92.52), but lacking statistical precision, with a large 95%CI, due to the small sample size. Individuals with visual impairment due to diabetes were 82% more likely to use proper insulin usage technique than those without visual complications from the disease (PR = 1.82; 95%CI: 1.00-3.30).

Other analysis

Observation of insulin preparation and application in the subsample: a total of 14 participants were observed preparing and administering insulin (four who used insulin vials and 10 who used insulin pens) (results not shown in tables). Of a possible 44 points for preparing insulin vials, those observed achieved a total of 11 points (25.0%). Among all those who used insulin pens, the participants obtained a total of 40 points (66.7%) for preparing insulin out of a possible 60 points. Regarding insulin administration, the proportion of proper procedures among all vial insulin users was 60.7% (17 out of 28 points), while among all the pen insulin users, the proportion was 71.3% (57 out of 80 points). The proportion of proper preparation and application was 53.3% among vial insulin users and 61.3% among pen insulin users (p = 0.253).

Discussion     

This study showed that the participants properly followed only two-thirds of the recommended steps for insulin use. The most frequent errors occurred during the preparation and administration of insulin. Regarding storage, all unopened insulin and most insulin in use were kept in the refrigerator. The average temperature of the environment in which the insulin in use was stored outside the refrigerator was within the recommended range of 15 to 30°C, even considering that data collection took place at summertime in Brazil. Other studies that investigated insulin storage found similar results [14-22].

Although a great proportion of participants spontaneously reported homogenizing the mixture before use, only a minority did so at least 20 or 10 times, as recommended, respectively, for a vial and a pen user. Given that failure to homogenize properly can result in leaving insulin crystals out of suspension, inadequate homogenization can alter insulin concentration, resulting in unpredictable clinical responses [12]. This can result in injecting less or more insulin than desired and in varying the injected concentration. Our results are consistent with those of other studies that evaluated preparation procedures and identified flaws in several steps [9,15,17,18,22-24]. The least mentioned application procedure in our study was keeping the needle in the tissue after application, even after the prompting question. Previous studies found similar results [9,15,18,24]. Leaving the needle in the subcutaneous tissue after injection helps ensure the entire dose is administered and prevents insulin reflux and leakage through the needle after removal [12].

Less than half of the participants reported aseptic conditions at the injection site. Other authors reported similar results [9,15,19,21]. One study showed that five seconds of skin cleansing with 70% alcohol reduced the bacterial count by 82%-91%, and that over three to five months of observation, more than 1,700 insulin injections were administered without skin preparation. Nonetheless, no signs of local or systemic infection were observed, thus indicating that although routine skin preparation with alcohol before injection significantly reduces the bacterial count, it may not be necessary to prevent infections at injection sites [25].

Only a minority reported pinching the skin before applying insulin. Injections with 8 mm needles, or even 4-6 mm needles in individuals with insufficient subcutaneous tissue, require skin pinching to avoid intramuscular injection [12]. It is important to note that we did not investigate the needle size or the participants' body mass index to assess the adequacy of this practice. Most participants reported rotating the injection site and using the appropriate angle of application. In other studies, the proportion of participants who followed these two steps varied. In Brazil, the proportion of participants who rotated the injection site varied from 18.2% [22] to 88.7% [15]. In European countries, reported rotation ranged from 38% [26] to approximately 70% [21,23]. Regarding the angle, some authors found low knowledge or practice [14,17,20], while others found that participants had knowledge of and/or applied the injection at an appropriate angle [9,18,22]. Most participants reported reusing the needle a maximum of eight times, as recommended, provided that it was not contaminated or caused discomfort during administration. One study conducted in Brazil found that 94.6% of participants reused needles one to 20 times after the initial use [27], while the average number of reuses in European countries was 3.6 times [24].

The current study identified several factors associated with the most appropriate usage technique: brown, black, Asian, or indigenous ethnicity; a higher level of education; reported alcohol consumption in the last month; older age at disease diagnosis; and the presence of visual impairment due to diabetes. No other studies were found in the literature that evaluated the association between appropriate insulin use and participants' color/ethnicity. We found no difference in ethnicity in the proportion of individuals who received guidance from professionals on insulin use that could justify a higher prevalence of appropriate use in one group than in another. However, it should be noted that the number of brown, black, Asian, and indigenous participants was small, which may have led to a spurious association with more appropriate insulin use. In the current study, most brown, black, Asian, and indigenous individuals used insulin pens. The finding regarding higher levels of education was also reported in several other studies [14,16,17,21,22,28]. Low education limits an individual's ability to understand guidelines and develop self-care skills. Impaired reading, writing, and comprehension skills negatively affect patients' adherence to treatment. Participants who reported consuming alcoholic beverages were more likely to use insulin appropriately. Since data collection took place in January and the question referred to the previous month, it is possible that participants reported consuming alcohol during end-of-year celebrations, and that the response does not represent the usual behavior. In our study, an older age at diagnosis was associated with shorter disease duration (p < 0.001). Other studies found no association between time since diagnosis and proper insulin usage technique, but the median number of errors decreased with time of insulin use and time since diagnosis, though not significantly [15,28]. In our study, the presence of visual impairment was associated with better insulin usage technique; however, two other studies found an inverse relationship [18,28]. It is important to note that 31 of the 61 individuals with visual impairment due to diabetes used pen injectors, and among 19 of them insulin was administered by another person rather than by the individual themselves. The International Diabetes Management Practice Survey of 2,475 patients with type 2 diabetes (43.3% of whom used insulin) from thirteen countries in Africa, the Middle East, Europe, Latin America, and Asia identified older age, a longer disease duration, having received diabetes education, and adherence to a healthy diet and exercise plan as factors associated with drug treatment adherence, including insulin usage technique [29].

Strengths and limitations: one strength of the study was the investigation of the three dimensions that comprise proper insulin use (storage, preparation, and administration), since most studies focus mainly on insulin administration. Additionally, conducting the study in participants' homes allowed for the collection of data, such as temperature verification and insulin storage location, which would otherwise be impractical or less reliable if the study were conducted in health facilities. The study with the subsample, corroborating the reported results, and the use of adjusted analysis are additional strengths of the study, given that most of the studies found in the literature were descriptive in nature.

On the other hand, the main limitation of the study is its small sample size, which limited the power to identify potentially statistically significant associations. Another limitation stems from the fact that the information was obtained through self-reporting. To address this issue, participants were asked to demonstrate where they stored their insulin, and a sub-study was conducted. Nonetheless, neither the location (shelf or door of the refrigerator) nor the temperature inside the refrigerator was verified. Additionally, we inquired about injection site rotation but did not investigate lipodystrophy, which would provide evidence of adequate rotation. Furthermore, for the association analyses, proper insulin use was defined as the proportion of individuals who used the technique more than the median in the analyzed sample since none of the participants fully met the adequacy criteria. Consequently, the appropriate insulin use in our study should be interpreted with caution, as it does not reflect the users’ adherence to all the recommended steps for the correct insulin use. Additionally, we did not investigate individual blood glucose monitoring or the reasons for low adherence. Therefore, it is impossible to determine whether low adherence is due to ignorance, forgetfulness, or free choice.

Conclusion     

This study, planned to assess the storage, preparation, and application of insulin among patients with diabetes in Canguçu, a city in South Brazil, showed that several steps in insulin use are often overlooked, underscoring the importance of evaluating this technique during routine follow-ups visits of patients with diabetes.

Competing interests     

The authors declare no competing interests.

Authors' contributions     

Luma Aquino de Avila: conceptualization, methodology, data collection, and writing-original draft preparation; Priscila Echevarria: methodology and data curation; Paulo Avila Junior: data collection; Ina da Silva dos Santos: conceptualization, methodology, statistical analyses, writing- reviewing and editing. All the authors read and approved the final version of this manuscript.

Acknowledgments     

The authors would like to thank the Municipal Health Department of Canguçu and the people with diabetes who depend on insulin, without whom this study would not have been possible. Luma Aquino de Avila received support from the Catholic University of Pelotas through the Dom Antônio Zattera Scholarship. ISS receives support from the National Council for Scientific and Technological Development (CNPq).

Tables     

Table 1: preparation of neutral protamine hagedorn insulin in vials (N=44) and in pens (N=70) by patients who got insulin from the Municipal Pharmacy in Canguçu, Brazil

Table 2: insulin application with syringe and needle (N=44) and with pen (N=69) by patients who got insulin from the Municipal Pharmacy in Canguçu, Brazil

Table 3: crude and adjusted prevalence ratios (PRs) with 95% confidence intervals (95%CIs) for proper insulin use by patients who got insulin from the Municipal Pharmacy in Canguçu, Brazil

Annex     

Annex 1: supplementary materials (PDF- 335Kb)

References