Recruitment
In this small municipality, we identified 20 patients with diabetes type 2 of whom 12 were men and 8 women. Eleven of them did not meet the age inclusion criterion so we asked 9 eligible patients to participate. Two of them did not want to participate due to severe illness (cancer, mental illness) and one felt bound and stressed by the thought of having different people from home care entering their home several times a day. Finally, six participants agreed to participate (Figure 1).
Feasibility for collecting sociodemographic and clinical data
We collected data from EPJ in home care on sociodemographic and clinical data and treatment goals (HbA1c). Complete sociodemographic and clinical data were available for all patients. Lack of data was identified for HbA1c. The case report forms for collecting sociodemographic and clinical data, meal diary and blood sugar measurements worked well.
Sample characteristics
The participants had a duration of diabetes of 15 years or more. We included four males and two females and the median age was 82 (75-88) years. Three persons lived alone. All persons were treated with insulin and five individuals used additional glucose-lowering medication. Median BMI was 26,1 (23,9-30,7).
Feasibility of CGM
The use of CGM among older home-dwelling people with diabetes was feasible. All patients completed all five days CGM measurement. Figure 2 displays an example of 5 days continuous glucose monitoring for one patient (patient 5).
Regarding practical use of CGM, one patient reported skin reaction (redness) and none reported pain. Furthermore, there were no technical errors, except for one sensor that fell off. Thus, the registration for this individual was delayed for one week. CGM data were obtained for a total of 29 days (one day missing for one individual). Calibration samples were collected without any specific challenges. All individuals performed three measurements of capillary blood glucose daily, three of them had help from home care nurses who performed the tests. For 27 of the 29 days the mean absolute relative difference (MARD) was ≤15%, while for the remaining 2 days the MARD was <20%. MARD is the average of the absolute error between paired samples, i.e. CGM values and corresponding capillary blood samples. Downloading the CareLink software to our computer was time consuming due to compatibility problems with more recent drivers.
Occurrence of hypoglycemia data
Regarding the CGM measurements, we collected 945-1140 sensor values between 3 and 22.2 mmol/L. TIR (3.9-8.3 mmol/L) varied between 0- 63%. In three individuals, blood glucose values were < 3.9 mmol / l, i.e. hypoglycemic episodes, with a duration of between 15-50 minutes (Table 1).
Table 1. Continuous Glucose Measure values in older people (≥75 years) with diabetes receiving home care (n=6).
|
Person 1
|
Person 2
|
Person 3
|
Person 4
|
Person 5
|
Person 6
|
Sensor values (n)
|
1126
|
1091
|
1154
|
1138
|
1148
|
945
|
Highest (mmol/L)
|
19.4
|
12.5
|
22.2
|
18.4
|
21.8
|
22.2
|
Lowest (mmol/L)
|
6.8
|
3
|
5.7
|
8.3
|
3.5
|
3.1
|
Average (mmol/L ± SD)
|
11.7±2.8
|
7.7±1.8
|
14.1±3.6
|
13 ±1.7
|
9.2 + 2.9
|
14.1±4.5
|
% within TIR1
|
10
|
63
|
7
|
0
|
45
|
9
|
% over 8,3 mmol/L
|
90
|
37
|
93
|
100
|
55
|
89
|
Under 3.9 mmol/L (n)
|
0
|
1
|
0
|
0
|
1
|
2
|
Duration under 3.9 mmol/L (min)
|
0
|
0:25
|
0
|
0
|
0:15
|
0:25/0:50
|
1Time-In-Range
Feasibility for questionnaires
Collecting questionnaire data during day 1 and 5 of the CGM period by two study nurses, assisted by the first author, was feasible. All patients completed the questionnaires. Some had difficulties understanding the meaning of some of the questions and we had to elaborate. As we assisted the patients in filling in the questionnaires some people with hearing impairment had more difficulties which made the data collection more time consuming. Median time spent on visit 1 was 32 (range 20-62) minutes versus 40 minutes on visit 2 (range 31-62) minutes (Figure 3). In total, we spent less time on the last persons that were included due to better routines. Time spent with persons living with others or having relatives present during the interview was higher. Regarding practical use of the MNA scale and IPLOS, interview manuals were used to support the questionnaire (available online) (20, 21, 23). The protocol initially contained four questionnaires, but the MMSE required extensive skills and training among home care personnel and could therefore not be performed.
Risk of hypoglycemia, malnutrition and level of functioning
The McKellar Risk Assessment tool was suitable to identify factors associated with hypoglycemia. All six individuals were at “very high risk of hypoglycemia” due to risk factors like kidney disease, stroke, swallow palsy, diarrhea, insulin treatment (n=5) and combination of insulin and sulfonylurea (n=1) in addition to polypharmacy (n=6). One person recently had a hypoglycemic event which required assistance. Regarding the McKellar risk survey, four out of six said that they did not notice any symptoms of hypoglycemia. Correspondingly, CGM measurements identified episodes of low blood glucose < 3.9 mmol/L in three of these four persons that did not notice any symptoms during the five days of data collection. In this questionnaire, one out of six reported HbA1c < 53 mmol/mmol and four out of six did not know their HbA1c level.
The Mini Nutrition Assessment was suitable to identify risk of malnutrition (Figure 3). Four out of six persons had risk of malnutrition (score 17-23 points). Some persons had difficulty swallowing, others had a physical dysfunction caused by stroke, poor appetite from depression, esophageal hernia and diarrhea and some forgot to eat because of their impaired cognitive functioning.
The Individual-based Statistics for Nursing and Care Services (IPLOS) was able to identify areas in physical dysfunctioning. The most common dysfunction areas were “common housekeeping”, “self-care” and “moving outdoors” (Figure 5). Common housekeeping gave the highest scores between 3-5 on (black) scale. Self-care (grey column) is the second highest score and several persons had disabilities (amputation of leg, leg wound, cognitive impairment/dementia) and were in need of medical attention and assistance. The third (shaded) column was moving outside which was difficult for all of them without the use of aids and assistance from others.