Patient and physician preferences for type 2 diabetes medications: a systematic review

Abstract

Background

There are several treatments to cure type 2 diabetes (T2D) and every one of them has certain attributes which is lead patients and specialists to have different preferences and select power. Therefore, we did this systematic study to evaluate patients̓ and physicians̓ preferences for type 2 diabetes medications by extracting attributes of anti-diabetic medications and identifying their relative importance.

Methods

We searched the PubMed, Ovid, Web of science, Scopus and Embase databases for articles which have been published on or before May 8th, 2018(The start time of the search in our study was May 8th, 2018).

Results

The searches identified 3346 studies, of which 17 (from 2009 to 2017) were included in the final synthesis and 27 attributes of type 2 anti-diabetic have been investigated. The most important attributes are changes of blood glucose and HbA1c level, hypoglycemia events, weight changes, gastrointestinal complications, cardiovascular effects, medicines cost, and administration mode and dosage of medicines.

Conclusion

physicians and patients prefer antidiabetics which is reduce blood glucose and HbA1c level effectively and have low side effects too (hypoglycemic event, cardiovascular and gastrointestinal). The effect of weight reduction, low cost, low dosing and low frequency of using. Health care providers, Specialist, and manufacturers should consider to these attributes in treatment process and marketing. It can increase adherence to management approaches, and reduce morbidity of patients with type 2 diabetes.

Introduction

The goal of treating type 2 diabetes is to reduce or to maintain the blood glucose to nearly normal levels (to the possible extent) which is lead to prevent of disease complications [1]. Despite of treatment benefits of anti-diabetic medicines, various studies have shown that less than 50% of patients achieve to these clinical goals. This impairment can be due to lack of compliance or adherence of patients with medication which should be according to clinical guidelines [2]. It seems that patient preferences are associated with increasing medicine compliance, clinical performance, treatment outcomes and reducing health care costs. On the other hand, the preferences and adherence of patients with diabetes mellitus depend on the patient’s concerns about the conditions and their belief about being diabetic. Other factors can be effective on patient’s preferences including blood glucose or hemoglobin A1c levels, insufficient understanding of the benefits and the side effects of medicines, the difficulty of following the recommended lifestyle, the cost of medications and health care, confusion about medicine usage, the psychological problems which are associated with managing chronic diseases [3–5].

There are a variety of oral and injectable anti-diabetic medications for treating T2D. Metformin is a preferable medicine for initiating therapy. If hemoglobin A1c levels do not reach the target level after three months of using metformin, the regimens of two-medicine, three-medicine, or four medicines will start. Therefore, a combination of metformin with the sulphonylurea and/or Thiazolidinedione and /or DPP-4 and/or SGLT2 and /or GLP-1 receptor agonist and/or insulin (basal) is prescribed [6]. Although the choice of medication or second-line medications is depend on the patient’s condition, there is no consensus between the specialists on the preferred pharmaceutical composition. In general, physicians consider: effectiveness, risk of hypoglycemia, renal effects, weight change, cardiovascular effects, oral/ subcutaneous and cost, and self-preferences (additional consideration) [6]. Therefore, physicians’ preferences can be various about diabetic medications.

However, physicians’ preferences can be very valuable because of logical and rational [7], studies have shown that the physicians’ preferences and patients’ preferences about treatments can be different. For example, there is a difference between physicians and patients view in terms of risk tolerance in the exchange between improvement and treatment efficacy as patients have lower tolerances. Also, patients appreciate the quality of life and health-related social situations more than physicians [8].

Considering that patient’ preferences for treatment adherence and the preferences of diabetes specialists for optimal medicine administration are important factors for controlling diabetes. Therefore, it is important to pay attention to patients̓ or physicians̓ preferences for the improvement and control of type 2 diabetes. In this regard, several studies have been conducted in different countries using discrete choice experiments (DCE) for assessment patients̓ or physicians̓ preferences [9–11].

The DCE is one of the best ways to identify the preferences of people in the community and valuation health Policy [12]. A DCE is a quantitative method increasingly used in healthcare to elicit preferences from participants (patients, payers, commissioners) without directly asking them to state their preferred options [13]. In fact, DCE is based on economic theory that assumes people can prefer good products and services rather than other services and products [12, 14]. Using the discrete choice experiments (DCEs), one can measure the preferences and values of individuals regarding the treatment regimens of diabetes [15].

As mentioned, the patients̓ or physicians̓ preferences play an important role in controlling and improving ddiabetes. On the other hand, the patients’ and physicians’ preferences are not aligned in some characteristic. So, we conformed this systematic review in order to assess patients̓ or physicians̓ preferences about anti-diabetic medicines and extract attributes of anti-diabetic medicines and their relative importance.

Materials and methods

Study design

In order to determine patients’ and physicians’ preferences regarding to diabetes medications (oral or injectable) for treating adults with T2D, we conducted a systematic review of published studies describing patient and physician preferences by using discrete choice method. A written study protocol was prepared in accordance with the MOOSE statements (Fig. 1) [16, 17], and the review was registered with PROSPERO (systematic review record CRD42018093228).

Fig. 1.

MOOSE flow diagram of observational studies included in the systematic review

Data sources and searches

We searched the PubMed, Ovid, Web of science, Scopus and Embase databases for articles published on or before May 8th, 2018 (The start time of the search in our study was May 8th, 2018). We used following terms: (“preferences” OR “revealed preference” OR “stated preference” OR “patient preference” OR “patient preferences” OR “physician preferences” OR “physician preference” OR “willingness to pay” OR” willingness to accept” OR “choice modeling” OR “discrete choice” OR “choice experiment” OR “discrete choice experiment” OR “contingent valuation”) AND (diabetes OR diabetics OR diabetes mellitus OR T2D). The keywords were limited by specialized perspective, for concentrating on the objective of the study and preventing diversity. So it could be mentioned as a limitation of study.

Study selection and data extraction

We included studies if they assessed patients̓ or physicians̓ preferences regarding diabetes medications (oral or injectable) for treating adults with T2D and if they studied Preferences using discrete choice experiments.

We excluded articles if they weren’t written in English, did not assess physicians’ treatment preferences, did not assess patients’ treatment preferences, or if they included only 18 years old patients, or if they contained no original data (e.g., review, commentary, editorial, or meeting abstract). In addition, we only assessed preferences for treating diabetes complications or comorbid medical conditions, only included non-medication diabetes treatment, and only assessed patients’ satisfaction with or adherence to treatment. After elimination of duplicates, two authors independently ruling out unrelated reports based on inclusion and exclusion criteria screened the title and abstract of retrieved articles. First, the results screened by their titles and irrelevant results excluded. Second, the abstracts of the selected results reviewed to eliminate conference abstracts, practice guideline, editorial letters, or reviews. Third, the full text of the chosen studies reviewed separately. Then, the two authors held a face-to-face meeting to compare their results with each other. Differences among opinions resolved by consensus and if necessary, a third author was involved for prevention selection bias. At least two reviewers reached consensus on study inclusion. In the next step, the full texts of the included articles were double- checked and evaluated in details in order to eliminate some other reports according to the exclusion/inclusion criteria (Fig. 1).

For Quality Assessment of studies, two reviewers used a checklist of the quality of cross-sectional studies (AXIS) [18]. Likewise, the risk of bias within each included study was assessed based on the unclear objective, unjustified sample size, undetermined statistical significance, incomplete outcome data, and non-described responders with ratings of “low risk of bias,” “high risk of bias,” and “unclear” (uncertain risk of bias). We didn’t see high risk for bias types (selection, reporting, performance, and attrition) in 17 included articles.

Finally, the requested data were extracted and summarized in the data sheet. The relevant data including the name of first author, publication date, place of the study, study design, sample size, and outcomes (attributes and their relative importance in viewpoints patients and physicians) were presented in the tables. After primary analysis, it was clear that because of different metrics of the same outcome, dissimilar study designs, non-normality of data, we were not able to conduct a meta-analysis or a quantitative analysis. Therefore, a qualitative analysis of the data conducted.

Data analysis

The findings of pertinent articles were extracted and were evaluated qualitatively and descriptively. We extracted attributes of the included studies, the number of times the attributes of anti-diabetic medicines reported, and the relative important attributes of anti-diabetic medicines, which are associated with patients’ or physicians’ preferences in T2D treatment based on discrete choice experiment method.

Results

Totally, 17 studies (that have been conducted on attributes of type 2 anti-diabetic medicines using method of discrete choice experiments) included. These studies have been conducted in Europe and European countries (Germany, Spain, Portugal, Swiss, England, Denmark and Sweden) (12studies) [9, 10, 15, 19–27], the United States (7 studies) [10, 11, 27–31], Brazil [9], South Africa [10], and Canada [9]. Most studies have evaluated the attributes of medicines from the patients̓ point of view (16 studies) [9, 10, 15, 19–31]. Only in three studies, these attributes surveyed from the perspective of physicians and endocrinologists [11, 20, 22]. In addition, most studies considered both oral and injection medicines (10 studies) [9, 15, 19, 20, 22, 23, 26–29]. The attributes of oral and injection (insulin and GLP1) medications had been surveyed singly in two [15, 21] and five studies respectively [11, 21, 25, 30, 31] (Table 1).

Table 1.

Attributes of the studies entered into the study

Row	Author (es)	Year	Country	Type of study	N	subjects	Drug type	Main Results
1	Carol Mansfield et al. [20]	2017	Germany and Spain	cross-sectional	875	Patients and physicians	Oral and injection	Risk of GI problems was most important to German respondents. MAB analysis found that respondents would require a 56 percentage point increase in the probability of reaching their HbA1c target to offset a change from 0% to 30% risk of GI problems. For Spanish respondents, mode of administration was the most important attribute.
2	Michael D Feher [10]	2016	South Africa, North America, Europe	cross-sectional	3758	patients	Injection	In the three regions, monthly willingness to pay was US$116, US$74, and US$92, respectively, for a 1%-point decrease in HbA1c; US$99, US$80, and US$104 for one less major hypoglycemic event per year; and US$64, US$37 and US$60 for a 3 kg weight decrease. To avoid pre-injection preparation of insulin, the respective values were US$47, US$18, and US$37, and US$25, US$25, and US$24 for one less injection per day.
3	Axel Mu¨hlbacher Susanne Bethge [21]	2016	Germany	cross-sectional	626	patients	Oral	The attributes, “additional healthy life years” (coefficient 0.458) or “additional costs” (coefficient 0.420), were in the middle rank. The side effects, risk of genital infection (coefficient 0.301), risk of gastrointestinal problems (coefficient 0.296), and risk of urinary tract infection (coefficient 0.241) followed in this respective order. Possible weight change (coefficient 0.047) was of less importance (last rank) to the patients in this evaluation.
4	Emuella M. Flood BA [28]	2016	US	cross-sectional	167	Patients	Oral and injection	The most influential attributes driving preferences were regimen, risk of diarrhea, weight change, risk of hypoglycemia, and efficacy.
5	A. Brett Hauber et al. [29]	2015	US	cross-sectional	923	patients	Oral and injection	Reducing dosing frequency was statistically significantly important to patients; however, it was relatively less important than medication cost or clinical outcomes. On average, patients preferred once-weekly to once-daily dosing.
6	Carlos Morillas et al. [22]	2015	Spain and Portugal	cross-sectional	551	Patients and physicians	Oral and injection	Patients valued avoiding a gain in bodyweight of 3 kg/6 months (WTP: €68.14(the most, followed by avoiding one hypoglycemic event/month (WTP: €54.80). Physicians valued avoiding one hypoglycemia/week (WTP: €287.18 the most, followed by avoiding a 3 kg/6 months gain in bodyweight and decreasing cardiovascular risk (WTP: €166.87 and €154.30, respectively). Physicians and patients were willing to pay €125.92 and €24.28, respectively, to avoid a 1% increase in glycated hemoglobin, and €143.30 and €42.74 to avoid nausea.
7	A. Brett Hauber et al. [30]	2015	us	cross-sectional	643	patients	Injection	Changing injection frequency from daily to weekly (independent of the effect of injection frequency on preferences for other attributes) was the most important predictor of treatment choice. Switching from a longer and thicker needle to a shorter and thinner needle and eliminating injection-site reactions were also statistically significant predictors of device choice (P < 0.05).
8	Christine Poulos et al. [11]	2013	us	cross-sectional	404	physicians and endocrinologists	Injection	The most important attributes to physicians were glucose control, risk of a fatal myocardial infarction (MI), and weight change. For US physicians, glucose control was about twice as important as gastrointestinal side effects, 5 times more important than changes in depression symptoms, and 20 times more important than liver monitoring. For UK physicians, reduction in MI risk was about 1.5 times more important than glucose control, 2.5 times more important than gastrointestinal side effects, and 10 times more important than liver monitoring requirements. Preferences were similar among physicians in the US and the UK and among GPs and endocrinologists.
9	H. L. Gelhorn et al. [23]	2013	Germany	cross-sectional	100	patients	Oral and injection	The total preference weight 9 PW) and relative importance (RI) were highest for four of the seven attributes: hypoglycaemic events (PW = 1.98; RI = 24.7%), weight change (PW = 1.65; RI = 20.6%), gastrointestinal/nausea side effects (PW = 1.49; RI = 18.6%) and efficacy (PW = 1.44; RI = 18.0%).
10	A.F. Mohamed et al. [24]	2013	Germany and Swiss	cross-sectional	383	Patients	Oral	For both countries, weight gain was the most important attribute, followed by blood glucose control. Avoiding a 5-kg weight gain was 1.5 times more important in Sweden and 2.3 times more important in Germany than achieving moderate blood glucose control, thereby, suggesting that blood glucose control is relatively more important to Swedish than to German patients.
11	MetteBøgelund et al. [15]	2011	Denmark	cross-sectional	270	Patients	Oral and injection	Patients placed the most value on losing weight and were willing to pay the most to lose 6 kg of weight. Loss of 3 kg of weight was the next highly valued, followed by dropping one percentage point in HbA1c level. Avoidance of nausea and a reduction in hypoglycaemic events from two per month to none was also highly valued. Patients were willing to accept one injection per day if they, for instance, simultaneously lost 1.4 kg.
12	Andrew Lloyd et al. [25]	2011	England	cross-sectional	252	Patients	Injection	The highest WTP values in participants with T1DM were avoiding 2-kg or 4-kg weight gain (£29 and £58, respectively), avoiding major difficulties with the injection device (£49), increasing the number of days per week when blood glucose levels are in the target range from 2 to 6 (£40), reducing the number of daily injections from 3 to 1 (£39), and avoiding nighttime hypoglycemia (£33). In participants with T2DM, similar factors had the highest WTP
13	Michael Polster et al. [31]	2010	US	cross-sectional	382	Patients	Injection	Efficacy measured by hemoglobin A1C was the most important attribute, followed by nausea, hypoglycemia, and dosing schedule.
14	J. Jendle et al. [26]	2010	Sweden	cross-sectional	461	Patients	Oral and injection	Patients wanting to reduce the number of hypoglycaemic events from three per month to none were willing to pay SEK 419 per month. Patients valued a 1 percentage point reduction in HbA1c at SEK 414 per month. Participants preferred taking tablets to injections and required a compensation of SEK 376 to accept one injection/day. Injections independent of meals were preferred to injections with meals (WTP: SEK 140 per month).
15	Franz Porzsolt et al. [19]	2010	Germany	cross-sectional	887	Patients	Oral and injection	The preference weights confirm that the reduction of body weight is at least as important for patients especially obese patients - and physicians as the reduction of an elevated HbA1c. Original products were preferred by patients while general practitioners preferred generic products.
16	A. B. Hauber et al. [27]	2009	US and England	cross-sectional	407	Patients	Oral and injection	Glucose control was the most important medication feature, followed by medication-related cardiovascular risk and weight gain, respectively. Water retention was not important to patients. Weight gain and cardiovascular risk had significant negative effects on likely medication adherence.
17	CamilaGuimara˜ es et al. [9]	2009	Canada and Brazil	cross-sectional	652	patients	Oral and injection	Glucose control was the most valued attribute by all socioeconomic groups, while route of insulin delivery was not as important. Patients with higher incomes were willing to pay significantly more for better glucose control and to avoid adverse events compared to lower income groups. In addition, they were willing to pay more for an oral short-acting insulin ($Can 71.65) compared to the low income group ($Can 9.85).

Overall, 27attributes of anti-diabetic medications were extracted. The attributes that evaluated in more studies included: blood glucose and hemoglobin A1c levels (16 studies) [10, 14–25, 27–29], weight changes (16 studies) [10, 14–25, 27, 28], hypoglycemia (15 studies) [10, 15–25, 27, 29], dosing frequency (11 studies) [10, 15, 16, 21, 22, 24–29]. Also other attributes that evaluated in more studies were: the gastro-intestinal problems (nausea, vomiting and diarrhea) (9 Studies) [10, 15, 17, 18, 20, 22, 23, 27, 29], cost of medications (more payment) (8studies) [10, 16–18, 20–22], risk of serious heart attack or stroke (7studies) [10, 15, 19, 23, 25, 27, 28], mode of administration (3 Studies) [9, 20, 26], blood sugar monitoring (4 studies) [10, 18, 22, 28] and genital tract infection and urinary tract (3 studies) [17, 19, 25] (Table 2).

Table 2.

Levels of attributes of anti-diabetic medicines in studies

Row	Variable	Levels				Total number of times reviewed in studies
		Two	Three	Four	Seven
1	Control blood sugar (HbA1c level)	2	14	0	0	16
2	Hypoglycemic events	3	6	5	0	14
3	Weight change	1	7	6	1	15
4	Dosing frequency	2	3	6	0	11
5	Risk of gastrointestinal problems	4	5	2	0	11
6	Additional payment per month	1	2	5	1	8
7	Risk of serious heart attack or stroke	4	2	1	0	7
8	Blood glucose monitoring	1	0	3	0	4
9	Mode of administration	2	1	0	0	3
10	Risk of urinary tract or genital infection	0	3	0	0	3
11	Type of product	1	0	0	0	1
12	Blood pressure	1	0	0	0	1
13	Injection device	1	0	0	0	1
14	Needle you use to inject	1	0	0	0	1
15	Need to store the medicine in a refrigerator	1	0	0	0	1
16	Changes in depression symptoms	0	1	0	0	1
17	Water retention	1	0	0	0	1
18	Ease of use of insulin device	0	0	1	0	1
19	Preparation of insulin before injection	1	0	0	0	1
20	Liver Damage	0	1	0	0	1
21	Pancreatitis	0	1	0	0	1
22	Lactic Acidosis	0	1	0	0	1
23	Malignancy	0	1	0	0	1
24	Additional healthy life years	0	1	0	0	1
25	Upper Respiratory Infection	1	0	0	0	1
26	Bumps or nodules around the injection site	1	0	0	0	1
27	Liver monitoring test	1	0	0	0	1

The levels of attributes

Results have shown that the main effects of anti-diabetics medications (HbA1c and fasting blood glucose) have been surveyed in three or two levels (fasting blood glucose levels: optimal, almost desirable and hemoglobin A1C levels: less than 1%, between 1% and 3%, and more than 3%) [10, 14–25, 27–29]. Type 2 anti-diabetic medicines increase or decrease weight or may not have effect on weight. This characteristic has been studied in two [29],three [9, 19–21, 24, 26, 28], four [10, 15, 22, 23, 26, 27]) and five [11] levels in the range of 6 kg weight gain to 6 kg weight loss in various studies. The average number of hypoglycemia events per a month evaluated in the two [10, 29, 31], three [9, 21, 23, 24, 26, 28], and four [10, 15, 20, 22, 27] (between zero and more than two cases). Gastrointestinal complications (including nausea, vomiting and diarrhea) studied at probability levels between 0% and 30%, or between non-emergence and emergence after three months. This characteristic studied in two [15, 22, 26, 31], three [11, 21, 23, 28, 29] or four [20, 27] levels. Risk of serious heart attack or stroke (as secondary effect) have been studied in two [20, 22, 23, 29], three [11, 28], and four [26] levels between zero and 3 %. Levels for blood glucose screening studied in two and four levels [15, 22, 26, 27] in the range between zero and three times a day. Also, The cost of paying more per a month have been studied in the range between $ 0 and $ 300 for better choice than the different levels of medicine attributes in eight studies [10, 15, 21, 22, 24–26, 29]. In studies, the levels of dosing frequency ranged between using two medicines twice a day and once a week (such as Bydureon). Most studies have divided this characteristic into four levels: one dose once a week, one dose once a day, two doses once a day, two doses twice a day (seven studies) [15, 20, 22, 26] (Table 2).

Regime (combination of mode of administration and dosing frequency) was shown in one article in seven levels [25]. Also, the Mode of administration (including: oral, injectable and inhalable) studied in two levels or three levels [9]. The urinary tract and genital tract infections as the side effects of diabetic medicines have been investigated at a probability level of zero to 20 % in three levels (three studies) [21, 23, 28]. Other attributes of type 2 anti-diabetic medicines investigated in three levels (6 studies) including: effects on malignancy [28], lactic acidosis [28], liver [28], pancreatic inflammation [28], respiratory infections [28], injection device [30], changes in depression symptoms [11], and healthy years of life [21]. Also, effects on adrenal glands, blood pressure [23], liver monitoring test [11], needle size [30], type of product [19], bumps or nodules around the injection site [30], storage time in the refrigerator [30], and preparation of insulin before injection [10] investigated in two levels. Ease of use of insulin device investigated in four levels [25]. These attributes considered in only one study (Table 2).

The most important attributes

Tables 3 and 4 shows the investigated attributes and their relative importance (By frequency and percentage). At Table 3, the frequency of each attributes and relative importance of them are listed (the attribute whether it was important (Yes) or less important (YN) significantly, or no important significantly (No)), from the viewpoints of patients and physicians which are collected by included articles. Table 4 shows the percentages of physicians and patients who considered attributes to be important significantly.

Table 3.

Attributes of anti-diabetic medicines and their importance (by frequency) in studies

Attributes
Studies	Control blood sugar (HbA1c)	Weight change	Hypoglycemic events	Risk of gastrointestinal problems	Dosing frequency	Additional payment per month	Risk of serious heart attack or stroke	Blood glucose monitoring	Mode of administration	Risk of urinary tract or genital infection	Blood pressure	Type of product	Injection device	Needle you use to inject the Medicine	Need to store the medicine in a refrigerator	Changes in depression symptoms	Water retention	Ease of use of insulin device	Preparation of insulin before injection	Liver Damage	Pancreatitis	Lactic Acidosis	Malignancy	Additional healthy life years	Upper Respiratory Infection	Bumps or nodules around the injection site	Liver monitoring test
1	Yesa	Nob	Yes	YNc	No	–	No	–	No	–	–		–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
2	Yes	Yes	Yes	–	YN	No	–	–	–	–	–	–	–	–	–	–	–	–	Yes	–	–	–	–	–	–	–	–
3	Yes	No	Yes	No	–	YN	–	–	–	No		–	–	–	–	–	–	–	–	–	–	–	–	No	–	–	–
4	Yes	Yes	Yes	Yes	Yes	–	Yes	–	–	No	–	–	–	–	–	–	–	–	–	No	No	No	No	–	Yes	–	–
5	–	–	YN	–	Yes	–	YN	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
6	Yes	Yes	Yes	YN	–	Yes	Yes	No	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
7	Yes	Yes	–	Yes	Yes	Yes	–	–	–	–	–	–	Yes	Yes	No	–	–	–	–	–	–	–	–	–	–	Yes	–
8	Yes	Yes	–	YN	YN	–	Yes	No	–	–	–	–	–	–	–	No	–	–	–	–	–	–	–	–	–	–	YN
9	Yes	Yes	Yes	–	–	–	Yes	–	–	Yes	Yes	–	–	–	–		–	–	–	–	–	–	–	–	–	–	–
10	Yes	Yes	Yes	Yes	–	Yes	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
11	Yes	Yes	Yes	Yes	YN	Yes	–	No	–	–	No	–	–	–	–	–	–	–	–	–		–	–	–	–	–	–
12	Yes	Yes	YN	–	Yes	YN	–	–	–	–	–	–	–	–	–	–	–	Yes	–	–	–	–	–	–	–	–	–
13	Yes	–	Yes	Yes	Yes	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
14	YN	Yes	Yes	YN	Yes	Yes	Yes	No	Yes	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
15	Yes	Yes	–	–	–	–	–	–	–	–	–	Yes	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
16	Yes	Yes	YN	Yes	–	–	–	–	–	–	–	–	–	–	–	–	No	–	–	–	–	–	–	–	–	–	–
17	Yes	Yes	Yes	–	Yes	–	–		Yes	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–

a: yes = important attribute in study, b: No: not important, c: YN: less important

Table 4.

Attributes of anti-diabetic medicines and their relative importance (by percentage) in studies

Attributes
Perspective	Control blood sugar (HbA1c)	Hypoglycemic events	Weight change	Dosing frequency	Risk of gastrointestinal problems	Risk of serious heart attack or stroke	Additional payment per month	Mode of administration	Type of product	Preparation of insulin before injection	Injection device	Bumps or nodules around the injection site	Needle you use to inject the Medicine	Need to store the medicine in a refrigerator	Ease of use of insulin device	Upper Respiratory Infection	Blood pressure	Risk of urinary tract or genital infection	Blood glucose monitoring	Liver Damage	Pancreatitis	Lactic Acidosis	Malignancy	Additional healthy life years	Changes in depression symptoms	Water retention	Liver monitoring test
patients %*(n)**	100% (n = 9732)	100% (n = 8004)	84% (n = 9811)	79% (n = 4355)	78% (n = 2878)	68% (n = 2747)	31% (n = 5515)	59% (n = 2168)	100% (n = 887)	100% (n = 758)	100% (n = 643)	100% (n = 643)	100% (n = 643)	0% (n = 643)	100% (n = 252)	100% (n = 167)	27% (n = 370)	11% (n = 893)	0% (n = 731)	0% (n = 167)	0% (n = 167)	0% (n = 167)	0% (n = 167)	0% (n = 626)	–	0% (n = 407)	–
Physicians %*(n)**	100%(n = 625)	100% (n = 221)	100% (n = 625)	–	–	100% (n = 625)	100% (n = 221)	–	–	–	–	–	–	–	–	–	–	–	0% (n = 625)	–	–	–	–	–	0% (n = 404)	–	–

*: The percentage of people who attribute was important from their perspective. **n = Total respondents

Hemoglobin A1C and fasting blood glucose studied in 16 article and they considered less important except for one study [26]. weight changes and Hypoglycemia events investigated in 15 and 14 studies as they were not important or less important in two [20, 21] and three [23, 25, 28] studies, respectively. The Risk of serious heart attack or stroke and gastrointestinal complications surveyed in 7 and 11 studies, respectively. These variables were not important or less important in one [15, 20, 29] and five [11, 20–22, 26] studies, respectively. The administration mode and dosing frequency in surveying three and eleven studies were not important in 1 [20] and 4 [10, 11, 15, 20] studies, respectively. The characteristic of blood glucose monitoring has been studied in four articles [15, 22, 26, 27] and it was not significant. Additional payment per month was important in the 5 [15, 21, 22, 24, 26] of 8 studies for patients or physicians. The risk of urinary tract or genital infection was important in the one [23] of three studies. Blood pressure was important in one study [23] and it was not considered important in another studies [15]. The type of product (Brand or Generic) [19], injection device [30], Ease of use of insulin device [25], preparation of insulin before injection [10], upper respiratory infections [28], needle size [30], and bumps or nodules [30] which are emerged around the injection place were considered important in one study. Changes in depression symptoms [11], water retention [27], liver damage [25], pancreatitis [25], lactic acidosis [25], malignancy [28], liver monitoring test [28], need to store the medicine in a refrigerator [30], and additional years of healthy life [21] studied in only one study and they were not important in other studies (Table 3).

100% of physicians stated that control blood sugar (HbA1c), hypoglycemic events, weight change, risk of serious heart attack or stroke, and additional payment per a month were important attributes. Control blood sugar (HbA1c) and hypoglycemic events were important significantly among 100% patients. Weight change, dosing frequency, risk of gastrointestinal problems, risk of serious heart attack or stroke, mode of administration, additional payment per month and were important significantly among 84%, 79%, 78%, 68%, 59%, and31% patients, respectively. In this analysis we just considered attributes that repeated at least in two studies or more. (Table 4).

According to findings, the most important attributes of anti-diabetic medicines included changes in fasting blood glucose and HbA1c level, hypoglycemia events, weight changes, risk of serious heart attack or stroke, drug cost, and gastrointestinal complications (nausea, vomiting and diarrhea). In addition, the patients preferred the main therapeutic effect (decreasing blood glucose) on choosing administration mode and the effect of administration mode on changing their weight, and physicians preferred the therapeutic effect (decreasing blood glucose) on weight changes and the effect of weight changes and administrations mode.

Discussion

Based on findings, the medicines of type 2 anti-diabetic have 27 attributes from perspective of patients and physicians. Among these attributes, the most important attributes include changes in fasting blood glucose and HbA1c level, hypoglycemia events, weight changes, digestive problems (nausea, vomiting and diarrhea), risk of serious heart attack or stroke, and medicines cost from perspective of diabetic patients and physicians by frequency (attributes in included articles) and percentage (The percentage of people who attribute was important from their perspective.), respectively. Mode of administrations and frequency dose were important only for patients.

In general, these variables have been investigated between two and seven levels in different scenarios, although they usually examined in two or four levels. We didn’t find any study on the attributes of anti-diabetic medicines using the discrete choice experiments in the continent of Asia.

Decreasing blood glucose and HbA1c level are the most important attribute of diabetes medications. Patients and physicians prefer medication that have high efficacy in decreasing blood glucose and HbA1c level. HbA1c levels reflect overall glycemic exposure over the past 2–3 months and are determined by both fasting and postprandial plasma glucose (PPG) exposure [32]. Clinical trials have demonstrated that reducing HbA1c and blood glucose lessens the risk of macro- and microvascular complications for patients with T2D [33, 34]. Therefore, Main effectiveness of anti-diabetic drugs in reducing blood glucose and hemoglobin A1C levels is considered as the most important characteristic from perspective of patients and physicians. Also, Reducing HbA1c levels lessens the risk of death [35] and is sign for improvement in the disease, so patients and physicians prefer this characteristic to others.

Another important attributes is changing weight by using diabetes medication from patients and physicians perspectives. Anti-Diabetic medications can cause decreasing, increasing, or no changing weight of patients. Physicians and patients like decreasing effect of anti-diabetic drugs on the weight. Because obesity is one of the risk factors for diabetes and cardiovascular disease [36] And on the other hand, it influences on the quality of life [37, 38]. Attributes of Weight loss is consider as a key factor in the management of T2D and the reduction of mortality associated with the disease [39]. As a result, physicians and health care providers suggested that diabetic patients avoid in weight increasing [40]. This attribute have very important qua patients valued weight loss and the avoidance of weight gain more than reduction HbA1c at Mette Bøgelund et al. study [15].

Among the important complications of anti-diabetic drugs, hypoglycemia events are the most important attributes followed by gastrointestinal and cardiovascular complications. In fact, physicians and patients prefer medication that are not associated with hypoglycemic events, and gastrointestinal and cardiovascular problems. Hypoglycemia events are accompanied with feeling of weakness or hunger, drowsiness, cold sweating, palpitations, dizziness, sleeping sickness. if these symptoms continue, it will ultimately lead to anesthesia and death [41]. Therefore, it is very important for patients and physicians. These outcomes are not surprising, since educational programs for health care providers have created an awareness of the risk of hypoglycemia, and its association with an increase in cardiovascular events and its relation to other detrimental effects such as lessening of treatment satisfaction and, ultimately, adherence to treatments [42].

Gastrointestinal (GI) complications include nausea, diarrhea and vomiting, usually occurring in patients who have recently started oral anti-diabetic medicines, which usually do not occur after several times of usage [43]. GI complications are important for improving both diabetic care and quality of life of the affected patient [44]. Thus, it is logical that GI complications be important for patients and physicians. Diabetics usually have cardiovascular problems [10]. Even when glucose levels are under control, it greatly increases the risk of heart disease and stroke. Therefore, reducing cardiovascular risk using anti-diabetic medicines is considerable for patients and physicians [36, 45].

Treatment costs controlling diabetes are relatively high for patients [46, 47]. Diabetes and its complications bring about substantial economic loss to people with diabetes and their families, and to health systems and national economies but some research suggested that people are less impatient over health issues than financial issues as patients were willing to pay for improvements in their T2D treatments [26, 48]. Therefore, it seems logical that the payment attribute has the fifth place of importance for physicians and patients. Yet, in Morillas et al. study showed that physicians were willing to pay up to three times more than patients for almost all attributes with the exception of receiving oral antidiabetics instead of injected medications [22].

Mode of administrations and dosing frequency are two other important attributes of anti-diabetic from perspective Patients. They usually prefer taking tablets to injections, independent using of meals to using with meals, and low dose to high dose, although in Guimara˜es et al. study, low-income individuals preferred short-term usage of insulin to oral drugs. This can be due to the attitude of patients about the effects of the injection, and most of these patients are among the low-income people. Basically, physicians and patients preferred that disease should be controlled with low dose medicine. It increases compliance with the treatments [49, 50].

This study included all studies on the preferences of patients and physicians about oral and injectable medications that examined using the discrete choice experiments and were published in English. Also, it surveyed both patient’s and physician’s preferences. These are considered as strong point of our study. Limitations of the study included: Studies have investigated the preferences using a method other than the discrete choice experiments have not been included in this study and we reviewed the studies in descriptively and did not use the meta-analysis method to extract the coefficient of attributes. Due to the heterogeneity of the results of the included studies, we were unable to perform a meta-analysis. In addition, the keywords limited by expert opinions, for preventing diversity and concentrating on the objective of the study. Our systematic review study completes the previous review studies [41, 47] which have only investigated the preferences of patients or physician about injection medications.

Conclusions

In general, this study showed that physicians and patients prefer antidiabetics that be reducing blood glucose and HbA1c level effectively, in addition to, those have low side effects (hypoglycemic event, cardiovascular and gastrointestinal), the effect of weight reduction, low cost, low dosing and low frequency of using. The preferences of physicians and patients are the same about almost attributes of anti-diabetic medicines. Our review provides available information for the policymakers, health system managers, and clinical on the relative importance of determinants that likely influence their treatment decision makings which are related to medications for treating T2D. Also, it gives a guide about the management approaches that might be more acceptable to different patients. It can increase adherence to management approaches, reduce morbidity, and improve quality of life of patients with type 2 diabetes. Also, medication manufacturers for T2D should develop medications that at low doses have high efficacy in lowering hemoglobin and blood sugar, as well as reduce the weight and risk of cardiovascular problems. On the other hand, they rarely cause hyperglycemia events and gastrointestinal problems.

Compliance with ethical standards

Conflict of interest

Authors declare that they have no conflicts of interest.