Differences in the prescribing patterns of diabetes medications among primary healthcare centers in Region Uppsala

(1)

Differences in the prescribing

patterns of diabetes medications among primary healthcare centers in Region Uppsala

A cross-sectional register studie

Patrimoine Ruremesha

Degree Project in Social Pharmacy, 30 credits, Fall 2020

Examiner: Sofia Kälvemark Sporrong

Division of Pharmacoepidemiology and Social Pharmacy Department of Pharmacy

Faculty of Pharmacy Uppsala University

(2)

Differences in the prescribing patterns of diabetes

medications among primary healthcare centers in Region Uppsala

Abstract

Background: Type II diabetes is a common disease worldwide and several drug treatments are available. Some of the more recently approved drugs are DPP4-inhibitors, GLP analogues and SGLT2-inhibitors. Most of the Type II diabetes drugs (T2DMD) are prescribed by

primary care physicians. To ensure rational drug use it is important to follow up prescribing patterns to design strategies and interventions that can improve drug treatment, since

consequences of inappropriate drug use might be poor health outcomes and increased health costs. Aim: To study differences in the prescribing pattern of T2DMD among primary healthcare centres (PHC) at a macro level in Region Uppsala. Method: A cross-sectional study based on data collected from Region Uppsala’s data register. Data consisted of individuals over 25 years of age with at least one prescription of a diabetes drug from the Anatomic Therapeutic Chemical (ATC) group A10 prescribed within the period of January 2018 to June 2020. Results: There is a moderate difference in the prescribing patterns of T2DMD among different PHCs in Region Uppsala. Overall, a larger proportion of oral antidiabetic drugs (OAD) were prescribed compared to Insulins. Most PHC prescribed a larger proportion of long acting insulins than other Insulins. Among OAD, a larger proportion of Biguanide derivative where prescribed. DPP-4 inhibitors, Sulphonylureas and SGLT2- inhibitors were prescribed to almost the same extent. Conclusion: Overall, there is a minor difference in the prescribing patterns of T2DMD among different PHCs in Region Uppsala during the study period. Lately, prescribing of DPP4-inhibitors, SGLT2-inhibitors and long acting insulins have increased.

(3)

Populärvetenskaplig sammanfattning

I detta fördjupningsprojekt studerades förskrivningsmönstren av typ II diabetesläkemedel i Region Uppsala. Studieresultaten vissa skillnader i hur olika vårdcentraler i Region Uppsala förskriver typ II diabetesläkemedel. Från början var hypotesen att icke-insuliner förskrivs i större utsträckning än insuliner. Vilket visade sig stämma överens med resultaten. I studien jämförs också skillnaden mellan förskrivning av icke-insuliner som funnits längre på marknaden och de nyligen godkända (DPP4-hämmare och SGLT2-inhibitorer). Resultaten visar en tydlig ökning i förskrivning av de nya läkemedlen under studieperioden. En anledning till detta kan vara läkemedels positiva farmakologiska effekter. I studien jämförs också skillnaden i förskrivningen av de olika insuliner i Region Uppsala. Resultaten visar att långverkande insuliner förskrivs i större utsträckning i jämförelse med andra insuliner, vilket är inte enligt rekommendationslistan för förskrivning av typ II diabetesläkemedel

Studieresultaten kan användas för att fortsätta följa upp och förbättra kvalitén på

diabetesvården i Region Uppsala. Ytterligare studier skulle till exempel kunna utgå ifrån hela populationen med en typ 2 diabetesdiagnos för att studera utfall även för andra interventioner än läkemedel. Dessutom behövs fortfarande kunskap om det är kostnadseffektiv att fler nya typ 2 diabetesläkemedel används. Studien ger kunskap om förskrivningen av typ II

diabetesläkemedel i Region Uppsala. Studien kan användas som grund för dialog om förskrivningen av typ II diabetesläkemedel i regionen, så att denna optimeras och fler

individer får adekvat behandling. Detta arbete är viktigt eftersom typ II diabetes är en mycket vanlig sjukdom där prevalensen ökar.

Denna studie är en tvärsnittsstudie baserat på sekundärdata från Region Uppsalas

elektroniska medicinalt register och primärdata från två enkäter skickade till sjuksköterskor och läkare som jobbar på vårdcentraler med patienter med diabetes. Studien inkluderar data från de 26 offentliga vårdcentraler som hör till förvaltningen Nära, Vård och Hälsa i region Uppsala. Förskrivningsdata baseras på recept för ett diabetesläkemedel från Anatomic Therapeutic Chemical (ATC) gruppen A10, förskrivet mellan januari 2018-juni 2020.

(4)

Acknowledgements

I want to acknowledge my supervisors Lena Ring and Anna Ekman for good supervision throughout the degree project. An acknowledgment to professor Björn Wettermark for all the wise and much needed opinions of how to improve the degree project further. Finally, an acknowledgement to my wonderful family and friends for all the support, positive energy and well wishes.

(5)

List of abbreviations

ATC- anatomic therapeutic chemical CV-cardiovascular disease

CVD- cardiovascular diseases EHR-electronic health records HbA1c- Haemoglobin A1c OAD-oral antidiabetic drugs PHC-primary healthcare centres PTL -per thousand listed

T2DM- Type II diabetes mellitus

T2DMD-Type II diabetes mellitus drugs

(6)

1.Introduction

1.1 Prevalence of diabetes mellitus

Diabetes mellitus is one of the world’s most prevalent diseases with high mortality rate in people younger than 70 years. In 2019, an estimate of 465 million people in the world had diabetes mellitus and this number is expected to double in the next 25 years (1). Especially in low- and middle-income countries (2). Geographically, 12,2 % of the population in North Africa and the Middle East was documented of having diabetes mellitus in 2019, 11,1 % of the population in North America and the Caribbean’s and 6,3 % of the population in Europe (1).

The prevalence of diabetes mellitus in the Nordic countries differs where 8,3 % of the population in Denmark was documented of having diabetes in 2019 ;5,3 % of the population in Norway and 4,8 % of the population in Sweden (3). The latter accounting for 438 157 people registered in the National diabetes register (NDR) were 12 852 of them were registered in Region Uppsala 2019 (4).

1.2 Diabetes mellitus

Diabetes mellitus is a group of metabolic diseases characterized by high blood glucose (hyperglycaemia) which can be caused by defects in insulin secretion, inactivation of the insulin or both. There are several types of diabetes and the two main types are Type I – accounting for 5-10% – and Type II – accounting for 90-95%. Type I is caused by a total deficiency in insulin secretion whereas type II is caused either by resistance to insulin or abnormal glucose levels because of insufficient insulin production (5).

Individuals with diabetes suffer from complications that are linked to chronic

hyperglycaemia. This can lead to long-term damage and sometimes even collapse of organs like eyes, heart, nerves and blood vessels (5). Diabetes can cause blindness, kidney failure, heart attack and stroke (2). Patients with this disease also have a higher risk to develop atherosclerosis (calcified vessels), abnormal levels of lipoproteins and they tend to have hypertension (5).

(10)

2

1.3 Treatment and care

Part of the treatment for Type II diabetes mellitus (T2DM) involves change of diet and increased physical activity so that the patient lose weight and eventually get lower blood glucose levels (2). When lifestyle changes are not enough, or the disease is too severe at the time of diagnosis drug treatment is necessary. In Region Uppsala, physicians have a list of guidelines for rational, safe and cost-effective drug prescribing called the “Recommendation list” (6). The Recommendation list is produced by the drug committee in Region Uppsala in collaboration with different medical expertise. These types of list are produced in all Swedish Regions. The equivalent list in Stockholm county council is called– the “Wise-list (Kloka Listan)” and adherence to this list specific is widely studied. These lists are used to increase prescribing quality and cost-effectiveness (7). There are three different types of

Recommendation lists: one for children, one for adults and one for the elderly (8).

The Recommendation list for adults is usually updated every two years but can be updated more frequently if necessary. It is mainly intended for primary care; it is useful to other medical specialities as well as when prescribing drugs that are outside their expertise. The list is in accordance with the Swedish and international authority guidelines, so physicians can optimize their prescription of these drugs. All drugs in the list have been approved for a minimum of two years on the market before being included. The drug committee also considers other things such as: how the drug is administrated, how easy it is to use and if the drug is included in the national health coverage (8).

Furthermore, the Recommendation list also includes guidelines for improving health habits such as: tobacco- and alcohol use, lack of or insufficient physical activity and unhealthy diet.

These guidelines are necessary as they affect the general wellbeing of patients and lifestyle factors accounts for approximately 80% of all cardiovascular disease and stroke (8).

(11)

3

1.4 Glucose-lowering medications

There are several Type II diabetes mellitus drugs (T2DMD) in the recommendation list.

These drugs can be first-line drugs, second-line drugs and some belongs to the group “other drugs”. See appendix 1.1 for T2DMD in the recommendation list.

1.4.1 First-line drugs

1.4.1.1 Biguanide derivative (Metformin)

Metformin is the first-line drug for treatment of T2DM according to the Recommendation list and other clinical guideline list such as the Wise list (8,9). The drug should always be used when tolerated as a first-line drug for treatment of T2DM unless in treatment of the elderly and fragile that should be treated with insulin, according to the Recommendation list (8).

Metformin is an oral antidiabetic drug (OAD) with the ATC code A10BA02 that belongs to the drug group; Biguanide derivative (1,1-dimetylbiguanide hydrochloride) (10). The drug has been used in Europe since 1957 and is suggested to act by improving the bodies

sensitivity for insulin. Metformin gives good glucose control, has a good safety profile and is cost-effective. The most common side effect caused by Metformin is gastrointestinal

intolerance (9). The drug have also been reported of having a low risk to cause lactic acidosis in patients with normal kidney function, 3-10 cases in 100 000 people (6).

The goal with treatment of T2DM with Metformin is for the patient to get good glucose control. If Metformin gives insufficient glucose control then a drug in the second-line is more suitable for treatment of T2DM (6). For mor information about Metformin see appendix 1.2.

(12)

4 1.4.2 Second line drugs

1.4.2.1 Insulin (Insulin- lispro and glargine and Insuman Basal)

Insulins are the oldest diabetes drug that has been used since the 1920s by individuals with T1DM and those with T2DM (8). Some negative sides of these drugs is being hard to get stable glycemia control with these drugs and that they are injectional fluid administrated subcutaneous (11).

There are three Insulin for treatment of T2DM in the recommendation list. These are Insulin lispro, Isuman Basal and Insuline glargine. Insulin lispro is a fast-acting insulin that should be used in combination with medium acting insulin or long-lasting insulin. The drug has the ATC code A10AB04 (10) and was the first Insulin approved by the Food and Drug

Administration (FDA) (12). Insulin lispro is suitable for use by individuals with T2DM that has low endogenous insulin production (6). Insuman Basal is a medium acting insulin with the ATC code A10AC01 (10) that is one of the second-line drugs in the recommendation list for treatment of T2DM. The drug is suitable for use by patients with an eGFR <30 ml/min and the elderly and fragile (6). Insulin glargine is a long-acting insulin with the ATC code A10AE04 (10) and is a suitable option as treatment for patients that experience

hypoglycaemia during night-time (6). For mor information about Insulin lispro, Isuman Basal and Insuline glargine see appendix 1.2.

1.4.2.2 Sulphonylureas (Glipizide and Glimepiride)

First generation of Sulphonylureas (SUs) were introduced on the drug market in the 1950s (13,14). These drugs has the ATC code A10BB (10) and can be used as monotherapy or as combination therapy with Insulin or Metformin. They act by binding to the receptor on the membrane of SUR cells and inhibits sensitive ATP K+ influx channels on pancreatic beta- cells ,which stimulates insulin secretion (14). SUs are cost-effective, safe and effective in reducing blood glucose levels. The negative side of these drugs is that they cause

hypoglycemia and weight gain (14).

There are two SUs listed in the recommendation list as second line drugs for treatment of T2DM suitable for use by the elderly and fragile (6).These are Glipizide (Mindiab^®) with the ATC code A10BB07 and Glimepiride with the ATC code A10BB12 (6,10). For more

information about these drugs see appendix 1.2.

(13)

5 1.4.2.3 Meglitinide (Repaglinide)

Repaglinide with the ATC code A10BX02 belongs to the drug class meglitinides (10). It was released on the pharmaceutical market in 1997 (13) and acts by selective binding to the pancreatic beta-cell membrane that inhibits ATP sensitive channels leading to stimulation of insulin secretion. The drug has the same effect as SUs though it acts faster and is structurally different from SUs. Repaglinide is easy to use as the drug can be taken during a meal. It has a rapid and dose dependent release of insulin and has a lower risk of causing hypoglycemia compared to SUs. Repaglinide has the same safety profile for cardiovascular (CV) risk as other OAD but is more expensive and have a complicated treatment scheme. The drug is mostly suitable for use by patients where Metformin is contra inducive (15). Repaglinide is also suitable for use by patients with an eGFR <30 ml/min (6) as the elimination of the drug is not primarily renal (15). The drug should be used with caution by the elderly and fragile (6). For more information about Repaglinide see appendix 1.2.

1.4.3 Other medication

1.4.3.1 Glucagon-like peptide-1 receptor agonists (Liraglutide, Dulaglutide and Semaglutide)

Glucagon-like peptide-1 receptor agonist (GLP-1 analogues) were discovered in 1987 but it took up to ten years for, exenatide, the first drug in this class to be approved by the Food and Drug Administrations (FDA). Today there are several GLP-1 analogues on the

pharmaceutical market such as Liraglutide, Semaglutide and Dulaglutide (16) that acts by stimulating insulin excretion (8). GLP-1 analogues have several benefits, some are their positive effects on the cardio vascular (CV), the glucose metabolism and secretion (16,17).

The negative side of GLP-1 analogues are their short half-life and that they have been suggested of causing acute pancreatitis and pancreatic cancer (13).

There are three GLP-1 analogues in the recommendation list for treatment of T2DM (6).

These are Liraglutide (Victoza^®) with the ATC code A10BJ02, Dulaglutide (Trulicity^®) ATC code A10BJ05 and Semaglutide (Ozempic^®) with the ATC code A10BJ06 (10). These drugs can be used by patients with cardiovascular diseases (CVD) and those that suffer from obesity (BMI >35). Victoza ^® and Ozempic ^®can be used by patients with an eGFR up to 15 ml/min (6). For more information about Victoza^®, Trulicity^® and Ozempic^® see appendix 1.

2.

(14)

6 1.4.3.2 Sodium-glucose co-transporter-2 inhibitors (Dapagliflozin, Kanagliflozin and

Empagliflozin)

Sodium-glucose co-transporter-2 inhibitors (SGLT2-inhibitors) with the ATC code A10BK, were released on the Swedish market in the end of 2003 (10,18). These drugs acts by

inhibiting natrium-glucose transporter 2 (SGLT) that normally reabsorbs glucose during glomerular filtration (8). SGLT2- inhibitors have good pharmacokinetic attributes, positive effects on the CV and do not cause weight gain. The negative side of SGLT2-ihibitors is that they cause increased risk for urinary tract- and fungal infections (13,17).

There are three SGLT2-inhibitors in the recommendation list for treatment of T2DM (6).

These are Dapagliflozin (Forxiga^®) with the ATC code A10BK01, Kanagliflozin (Invokana^®) ATC code A10BK02 and Empagliflozin (Jardiance^®) with the ATC code A10BK03 (10).

These drugs should be prescribed to patients with CVD and/or heart failure and to obese patients (BMI >35) with T2DM (6). For more information about Forxiga^®, Invokana^® and Jardiance^® see appendix 1.2.

1.4.3.3 Dipeptidyl peptidase-4 inhibitors (Sitagliptin and Linagliptin)

Dipeptidyl peptidase-4 inhibitors (DPP-4 inhibitors) are novel OAD with the ATC code A10BH. These drugs act by inhibiting the enzyme, dipeptidylpeptdas-4 (DPP-4) that breaks down the incretin hormones glucagonlike peptide-1 (GLP-1) and glucose-dependent

insulinotropic polypeptide (GIP). These two hormones normally increase insulin -and glucagon secretion (8). The positive side of DPP4-inhibitors is that they achieve good glycaemic control and weight reduction without causing hypoglycaemia (13).

There are two DPP4-inhibitors in the recommendation list for treatment of T2DM (6). These are Sitagliptin (Januvia^®) with the ATC code A10BH01 and Linagliptin (Trajenta^®) with the ATC code A10BH05 (10). Both of these drugs are appropriate for use by patients with reduced kidney function (eGFR <30ml/min) and the elderly and fragile (6). For more information about Januvia^® and Trajenta^® see appendix 1.2.

(15)

7

1.5 Target values

The main purpose with treatment of T2DM with antidiabetic drugs is to get a stable glucose control so that the risk of hypo- or hyperglycaemia and other diabetes complications reduces.

Target value for stable glucose control depends on the patient and is therefore set

individually. The target value of HbA1c-level for healthy and newly diagnosed patients is

≤42 mmol/ mol. All patients should target HbA1c levels ≤ 52 mmol/mol except for patients with difficult hyperglycemia, tough micro-and macrovascular complications, other diseases or those that are elderly and fragile. These patients have a higher target value than ≤52 mmol/mol (8).

Target value HbA1c-levels for the elderly and fragile with low life expectance (≤ 5 years) can be up to 70 mmol/mol because the goal with drug treatment to this patient group is different.

Treatment of this group is focused on improving life quality (reduce tiredness and

maintaining ADL-function during hyperglycemia), make sure that they have a good nutrition (20-30 kcal/kg/day) and safe (reduced risk for hypoglycemia) lifestyle (8).

(16)

8

1.6 Variables affecting prescribing patterns

There is not enough theoretical information about physician’s decision-making on drug prescribing. One of the latest conceptual models, based on previous theoretical models, to understand physician’s decision- making on prescribing is by Murshid and Mohaidin. This model suggests that there are three main factors that affects physicians’ decisions-making on prescribing. These are marketing efforts, patients’ characteristics and factors related to pharmacists (19).

Marketing efforts that affects physicians’ decisions-making on prescribing is the availability of drug information, the brand of the drug, sales and how the drug is represented and its effectivity. The drugs representation and effectivity are affected by the drugs characteristics, cost and benefit of the drug and the physician’s persistence to the drug. Patients

characteristics that affects physicians’ decisions-making on prescribing is the patients need of the drug and their expectation. The latter is affected by the drugs characteristics, cost and benefit of the drug and the physician’s persistence to the drug. Factors related to the pharmacists that affects physicians’ decisions-making on prescribing is the pharmacist’s knowledge and the collaboration between the prescribing physicians and the pharmacists.

These two factors are affected by the trustworthiness between the prescribing physicians and the pharmacists (19).

Why research prescribing patterns?

There are differences in the prescribing patterns within the same health institutions, countries, regions (7). This is of concern because inappropriate drug use increase side effects and cause avoidable morbidity and mortality, and is waste of resources (12). More research on

prescribing patterns might increase transparency in the medical field and the knowledge gained can be used to explain why prescribing patterns differ within similar health institutions. This in turn will make for a great foundation that the government can use to improve the quality of health care (10).

(17)

9

2.Aim and objectives

The aim of this project was to study differences in the prescribing pattern of drugs for treatment of Type II diabetes among primary healthcare centers (PHC) at a macro level in Region Uppsala, Sweden. The specific research questions were:

• Are there any differences in the prescribing patterns of Insulins compared to oral antidiabetic drugs (OAD), respectively in new-and old Insulins and new- and old OAD among different PHCs in Region Uppsala?

• Which factors related to the patient or the practice are associated with differences in the prescribing patterns of diabetes drugs?

-Is there any difference in prescribing of DPP-4 inhibitors, GLP analogues and SGLT2-inhibitors that depends on the size of PHC based on per 1000 listed patients (PTL), PHCs in urban-and rural areas and based on population mean income during the period January-June 2020?

• What are the experiences and perceptions of doctors and nurses regarding diabetes mediation and care at the PHC in Region Uppsala?

(18)

10

3. Methods

3.1 Study design & Data sources

This study is a cross-sectional study based on secondary data collected from Region

Uppsala’s electronic health records (EHR) and primary data collected through two surveys, sent to nurses and physicians that works with diabetes care, respectively. The EHR data were retrieved from a central data warehouse containing anonymized data from the EHR system Cosmic and data from a national primary care quality register (Nationell

PrimärvårdsKvalitet) including health data for people living in the region. Patients that have died were excluded from the data.

3.2 Population

The study included data from the 26 public PHCs that are run by Primary Care and Health in Region Uppsala. Prescribing data were collected for all individuals over 25 years of age with at least one prescription of a diabetes drug from the Anatomic Therapeutic Chemical (ATC) group A10, prescribed between January 2018 to June 2020.

23 out of 26 PHCs were included in this study though data from all the 26 PHC was used.

PHC was given anonymous names; A, B, C, D etc to preserve their integrity. These PHCs are located within the pink area in figure 1.

Figure 1: Map of Region Uppsala.

(19)

11

3.3 Data collection

Data from Region Uppsala EHR data warehouse were sent as an Excel file (Excel office 365) in an aggregated format where no individual patients could possibly be identified. This was further converted to a flat file with aggregated statistics on number of individuals prescribed different T2DMD per PHC. The data was divided in five-time periods; 2018 H1, 2018 H2, 2019 H1, 2019 H2 and 2020 H1. H1 describes the period January to June each year while H2 describes period July to December each year. The data set consisted of the following

variables; name of the PHC, number of listed individuals at the PHC, number of prescribed prescriptions for a drug with the ATC code A10 at the PHC, number of individuals that have been prescribed a prescription for a drug with the ATC code A10 at the PHC, the last time an individual had an appointment with physician or nurse, mean HbA1c-levels assessed at the PHC, patients age and the drug(s) prescribed under the ATC code A10.

Data of the experiences and perceptions of physicians and nurses regarding diabetes

mediation and care was collected through two surveys, sent to all the nurses and physicians employed at a PHC in Region Uppsala. The surveys were made by me and my two

supervisors and consist of closed- and open-ended questions (See appendix 2 for the complete surveys).

(20)

12

3.4 Data analysis

Analysis of PHC characteristics and proportion drugs prescribed with the ATC code A10 was conducted for the five periods: January- June 2018 (H1), July- December 2018 (H2),

January- June 2019 (H1), July- December 2019 (H2) and January 2020 to June 2020 (H1).

The proportion drugs prescribed can be more than 100 % since some individuals are prescribed more than one drug with the ATC code A10 during a period. In this section number of individuals listed means number of individuals listed at the PHC with a prescription for a drug with the ATC code A10.

3.4.1 PHC characteristics

PHC characteristics; number of prescribed prescriptions with the ATC code A10, number of individuals with a prescription with the ATC code A10, average age and average HbA1c are presented under point 4.1. The variables number of prescribed prescriptions with the ATC code A10 and number of individuals with a prescription with the ATC code A10 during January 2018-June 2020 where divided with per 1000 listed (PTL). The other two variables were summarized as average values.

3.4.2 Proportion drugs prescribed with the ATC code A10

The proportion of drug prescribed with the ATC code A10 was calculated by dividing number of individuals who got a prescription for a drug class e.g. oral antidiabetic medicines (A10B) with the number of individuals that was prescribed a drug with the ATC code A10 at a PHC in Region Uppsala.

In this study the size of the PHCs was based on the number of individuals listed. Analysing size dependent prescribing of DPP-4 inhibitors, GLP analogues and SGLT-inhibitors begun with dividing PHCs in three different categories: small, middle and big PHC. Small PHCs were defined as having <7000 individuals listed, middle PHCs <14 000 individuals listed, and big PHC >14 000 individuals listed. To calculate the proportion of e.g. DPP-4 inhibitors prescribed at e.g. small PHCs, the total number of individuals with a prescription for DPP-4 inhibitors at small PHC was divided with the total number of individuals listed at the small PHC. The same thing was done to calculate the proportion of GLP analogues and SGLT- inhibitors prescribed at middle and big sized PHCs.

(21)

13 Location depending prescribing of DPP-4 inhibitors, GLP analogues and SGLT-inhibitors begun with dividing PHCs in two different categories: countryside and urban area, depending on where the PHC is situated. In this study we choose to use our own definition of

countryside and urban area. To calculate the proportion of e.g. DPP-4 inhibitors prescribed at e.g. PHC located in the countryside, the total number of individuals with a prescription for DPP-4 inhibitors at the PHC located in the country side was divided with the total number of individuals listed at the PHC located in the country side. The same thing was done to

calculate the proportion of GLP analogues and SGLT-inhibitors prescribed at PHCs located in urban areas.

Income depending prescribing of DPP-4 inhibitors, GLP analogues and SGLT-inhibitors begun with dividing PHCs in seven different municipalities: Uppsala, Enköping, Östhammar, Heby, Knivsta, Älvekarleby and Tierp, located in Region Uppsala. Income was defined as the minimum wage in the municipalities before tax (24). To calculate the proportion of e.g. DPP- 4 inhibitors prescribed at e.g. PHC located in Uppsala, the total number of individuals with a prescription for DPP-4 inhibitors at the PHC located in Uppsala was divided with the total number of individuals listed at the PHC located in Uppsala. The same thing was done to calculate the proportion of GLP analogues and SGLT-inhibitors prescribed at PHCs located Enköping, Östhammar, Heby, Knivsta, Älvekarleby and Tierp.

3.5 Statistics

Descriptive statistical analysis was conducted in Excel 360. The proportion of drugs with the ATC code A10 prescribed was calculated. Pearson correlation coefficient was used to

determine the association between average HbA1c-levels and the proportion of DPP-4 inhibitors, GLP analogues and SGLT-inhibitors prescribed. Correlation was determined by plotting HbA1c-levels as the y-value against the proportion of the drug prescribed as the x- value. In theory, the strength of the association depends on how close the data is in the diagram which is determined by the coefficient (r). Low strength on the association has the value of 0,1 to 0,3 for a positive relationship and -0,1 to -0,3 for a negative relationship.

Medium strength on the association has the value of 0,3 to 0,5 for positive relationship respective -0,3 to -0,5 for a negative relationship. Association between 0,5 to 1 is of large strength for a positive relationship, -0,5 to -1 for a negative relationship (24).

(22)

14

3.6. Ethical considerations

As personal health data has been used in this study there are some ethical considerations that has been taken. All health data used is anonymous and has been used only for the planned analysis. All datasets have been handled according to regulations to minimize the risk of information getting into unauthorized hands. The benefits of using personal health data in this study will contribute to new knowledge about the prescribing patterns of diabetes drugs in Region Uppsala. This information can be used to improve treatment and care of patients with diabetes in the region.

(23)

15

4. Results

4.1 PHC characteristics from January 2018 to June 2020

PHC P prescribed the most drugs, 423 prescriptions PTL, with the ATC code A10 in Region Uppsala from January 2018 to June 2020 (Table 1). It is the PHC with most individuals, 124 PTL, with a prescription for a drug with the ATC code A10. PHC G prescribed the least drugs with the ATC code A10 (n=120 prescriptions PTL) and had the least individuals (n=26 PTL) during January 2018 to June 2020. On an average, individual was prescribed 2

prescriptions of a drug with the ATC code A10 at a PHC in Region Uppsala during this period.

The average age of individuals that were prescribed a drug with the ATC code A10 at a PHC in Region Uppsala during January 2018- June 2020 was 66 years. PHC X had individuals of highest average age (70 years) and PHC Q had individuals of lowest average age (61 years).

The average HbA1c of individuals that were prescribed a drug with the ATC code A10 at a PHC in Region Uppsala during January 2018- June 2020 was 60 mmol/mol. PHC A and PHC V had individuals with highest average HbA1c (61 mmol/mol) and PHC U had individuals with lowest average HbA1c (57 mmol/mol).

(24)

16 Table 1: PHC characteristics from January 2018 to June 2020: Name of the PHC, number of prescribed prescriptions with the ATC code A10 PTL, number of individuals with a prescription with the ATC code A10 PTL, the average age of individuals listed at the PHC and average HbA1c of individuals that got a prescription for a drug with the ATC code A10.

PHC

Prescribed prescriptions

*PTL

Individuals with a prescription

*PTL

Mean age (years)

HbA1c (mmol/mol)

PHC A 169 37 65 61

PHC B 182 39 65 60

PHC C 209 63 69 58

PHC D 372 71 67 60

PHC E 342 62 66 58

PHC F 155 35 68 60

PHC G 120 26 68 57

PHC H 176 37 65 60

PHC I 281 60 67 60

PCH J 329 99 64 65

PHC K 189 39 67 59

PHC L 346 67 67 59

PHC M 133 27 66 60

PHC N 177 49 65 60

PHC O 131 30 67 58

PHC P 423 124 66 60

PHC Q 214 45 61 64

PHC R 185 40 66 60

PHC S 130 35 64 63

PHC T 298 62 67 60

PHC U 308 69 69 57

PHC V 153 51 67 61

PHC X 292 61 70 58

*PTL= per thousand listed

(25)

17

4.2 Proportion insulin (A10A) and OAD (A10B) prescribed from January 2018 to June 2020

Figure 2. The proportion Insulins prescribed from January 2018 to June 2020: Max- and minimum proportion Insulins (A10A) prescribed at a PHC in Region Uppsala during 2018 H1 until 2020 H1. H1 stands for the period January to June, while H2 stands for the period July to December.

The maximum proportion of Insulins prescribed at a PHC in Region Uppsala during January 2018-June 2020 was 41 % Insulins (Figure 2). This was during 2018 H1, 2018 H2 and 2019 H1. The minimum proportion of Insulins prescribed at a PHC in Region Uppsala between January 2018- June 2020 were 12 % Insulins. This was during 2018 H1.

41% 41% 41%

34%

39%

12% 15% 17%

13% 14%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

2018 H1 2018 H2 2019 H1 2019 H2 2020 H1

Proportion prescriptions (%)

(26)

18 Figure 3. The proportion OAD prescribed from January 2018 to June 2020: Max- and minimum OAD (A10B) prescribed at a PHC in Region Uppsala during 2018 H1 until 2020 H1. H1 stands for the period January to June, while H2 stands for the period July to December.

The maximum proportion of OAD prescribed at a PHC in Region Uppsala during January 2018- June 2020 was 98 % OAD (Figure 3). This was during 2018 H1 and 2019 H2. The minimum proportion OAD prescribed at a PHC in Region Uppsala between January 2018- June 2020 was 81 % OAD. This was during 2018 H2.

98% 95% 94% 98% 97%

82% 81% 81% 83%

88%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

2018 H1 2018 H2 2019 H1 2019 H2 2020 H1

Proportion prescriptions(%)

(27)

19

4.3 Proportion drugs with the ATC code A10A prescribed

A large proportion of long acting insulin compared to fast-acting insulin and intermediate- acting insulin were prescribed at most PHC in Region Uppsala during 2018 H1 (See appendix 3.1). The minimum proportion of long acting insulin prescribed was 2 % at PHC G.

Maximum proportion long acting insulin prescribed was 28 % at PHC R. The median

proportion of long acting insulin prescribed at a PHC in Region Uppsala during 2018 H1 was 12 %.

Same as period 2018 H1, in 2020 H1 a large proportion of long acting insulin compared to fast-acting insulin and intermediate- acting insulin were prescribed at most PHC in Region Uppsala (See appendix 3.2). The minimum proportion long acting insulin prescribed was 6 % at PHC G. Maximum proportion long acting insulin prescribed was 28 % at PHC D. The median proportion of long acting insulin prescribed at a PHC in Region Uppsala during 2020 H1 was 15 %.

4.4 Proportion OAD (A10B) prescribed

During 2018 H1, a large proportion of Biguanide derivatives compared to other OAD were prescribed at all PHC in Region Uppsala (See appendix 4.1). The minimum proportion Biguanide derivatives prescribed was 55 % at PHC C. Maximum proportion Biguanide derivatives prescribed was 84 % at PHC G. The median proportion Biguanide derivatives prescribed at a PHC in Region Uppsala during 2018 H1 was 68 %.

Proportion DPP-4 inhibitors and SUs prescribed was almost the same during 2018 H1. The minimum proportion DPP-4 inhibitors prescribed was 10 % at PHC T. Maximum proportion DPP-4 inhibitors prescribed was 28 % at PHC X. The median proportion of DPP-4 inhibitors prescribed at a PHC in Region Uppsala during 2018 H1 was 19 %.

The minimum proportion SUs prescribed was 10 % at PHC I. Maximum proportion SUs prescribed was 25 % at PHC A and PHC V. The median proportion of SUs prescribed at a PHC in Region Uppsala during 2018 H1 was 16 %.

(28)

20 Same as during 2018 H1, in 2020 H1 a large proportion of Biguanide derivatives were

prescribed compared to other OAD (See appendix 4.2). The minimum proportion Biguanide derivatives prescribed was 57 % at PHC M. Maximum proportion Biguanide derivatives prescribed was 82 % at PHC G. The median proportion of Biguanide derivatives prescribed at a PHC in Region Uppsala during 2020 H1 was 65 %.

During 2020 H1, the proportion DPP-4 inhibitors and SGLT2-inhibitors prescribed were almost the same. The minimum proportion DPP-4 inhibitors prescribed was 11 % at PHC R.

Maximum proportion DPP-4 inhibitors prescribed was 26 % at PHC A and PHC N. The median proportion of DPP-4 inhibitors prescribed at a PHC in Region Uppsala during 2020 H1 was 21 %.

The minimum proportion SGLT2-inhibitors prescribed was 5 % individuals at PHC O.

Maximum proportion SGLT2-inhibitors prescribed was 33 % individuals at PHC L. The median proportion of SGLT2-inhibitors prescribed at a PHC in Region Uppsala during 2020 H1 was 17 %.

(29)

21

4.5 Difference in prescribing of DPP-4 inhibitors, GLP analogues and SGLT-inhibitors

Figure 4. PHC size dependent prescribing of DPP4-inhibitors, GLP analogues and SGLT2- inhibitors during 2020 H1: The proportion DPP4-inhibitors (A10BH), GLP analogues (A10BJ) and SGLT2-inhibitors (A10BK) prescribed at a small, middle or big PHC in Region Uppsala during 2020 H1. The size of the PHC is based on the number of individuals listed with a prescription for a drug with the ATC code A10. H1 stands for period January to June.

A large proportion of DPP-4 inhibitors, GLP analogues and SGLT-inhibitors were prescribed at small PHC compared to middle and big sized PHC in Region Uppsala during 2020 H1 (Figure 4). The difference in proportion of drugs prescribed at small, middle and big PHC was not great. 21 % DPP-4 inhibitors were prescribed at small PHC compared to 20 % at middle - and 18 % at the big sized PHC. 20 % GLP analogues was prescribed at small sized PHC compared to 17 % s at middle- and 20 % at big sized PHC. 18 % SGLT2-inhibitors was prescribed at small sized PHC compared to 12 % at middle- and 11 % at big sized PHC.

0%

5%

10%

15%

20%

25%

DPP-4 inhibitors (A10BH) GLP analogues (A10BJ) SGLT2-inhibitors (A10BK)

Proportion prescription (%)

Small (<7000) Middle (<14 000) Big PHC ( >14 000)

(30)

22 Figure 5. PHC location dependent prescribing of DPP4-inhibitors, GLP analogues and SGLT2- inhibitors during 2020 H1: The proportion of DPP4-inhibitors (A10BH), GLP analogues (A10BJ) and SGLT2-inhibitors (A10BK) prescribed at a PHC, located in the countryside or urban area, in Region Uppsala during 2020 H1. H1 stands for period January to June

A large proportion of DPP-4 inhibitors and GLP analogues were prescribed at PHCs situated in the countryside compared to those in urban in Region Uppsala during 2020 H1 (Figure 5).

The difference in proportion of drugs prescribed in these two areas was not great. 20 % DPP- 4 inhibitors were prescribed in the countryside compared to 19 % in the urban area.

Proportion GLP analogues prescribed in the countryside was 20 % compared to 16 % in urban areas. 12 % SGLT2-inhibitors were prescribed in the urban areas compared to 10 % in the countryside.

0%

5%

10%

15%

20%

25%

DPP-4 inhibitors (A10BH) GLP analogues (A10BJ) SGLT2-inhibitors (A10BK)

Countryside Urban area

(31)

23 Figure 6. PHC income dependent prescribing of DPP4-inhibitors, GLP analogues and SGLT2- inhibitors during 2020 H1: The proportion of DPP4-inhibitors (A10BH), GLP analogues (A10BJ) and SGLT2-inhibitors (A10BK) prescribed at a PHC in Uppsala, Enköping, Östhammar, Heby, Knivsta, Älvekarleby and Tierps municipality during 2020 H1. H1 stands for period January to June.

There was a slight difference in proportion DPP-4 inhibitors, GLP analogues and SGLT2- inhibitors prescribed based on the income before tax of the municipality situated in Region Uppsala during 2020 H1 (Figure 6). A large proportion of DPP4-inhibitors, 23 %, were prescribed in Heby and Knivsta municipality. The least proportion DPP4-inhibitors, 15 %, were prescribed in Älvkarleby municipality. A large proportion of GLP analogues, 27 %, were prescribed were prescribed in Tierp municipality. The least proportion GLP analogues, 11 %, were prescribed in Älvkarleby municipality. A large proportion of SGLT2-inhibitors, 19 %, were prescribed in Knivsta municipality. The least proportion SGLT2-inhibitors, 11 %, were prescribed in Östhammar municipality.

0%

5%

10%

15%

20%

25%

30%

Uppsala (26 683 kr)

Enköping (28 000 kr)

Östhammar (28 050 kr)

Heby (24 925 kr)

Knivsta (33 292 kr)

Älvekarleby (25 617 kr)

Tierp (25 000 kr)

DPP-4 inhibitors( A10BH) GLP analogues (A10BJ) SGLT2-inhibitors (A10BK)

(32)

24

4.6 Correlation of HbA1c with proportion DPP-4 inhibitors, GLP analogues and SGLT2-inhibitors prescribed

Figure 7. Correlation of HbA1c and proportion DPP4-inhibitors, GLP analogues and SGLT2- inhibitor prescribed during 2018 H1: The correlation of HbA1c and proportion of DPP4-inhibitors (A10BH), GLP analogues (A10BJ) and SGLT2-inhibitors (A10BK) prescribed from a PHC in Region Uppsala during 2018 H1. H1 stands for period January to June.

There was a general low association between HbA1c and proportion DPP-4 inhibitors, GLP analogues and SGLT2-inhibitors that were prescribed at a PHC in Region Uppsala during 2018 H1 (Figure 7).The association between HbA1c and proportion DPP-4 inhibitors

prescribed had the strongest negative correlation with HbA1c, R²= -0,3542, compared to the other two drug groups during this period. Correlation of HbA1c and proportion GLP

analogues prescribed was R²=0,1189 while that of SGLT2-inhibitors was R² =0,0003.

R² = 0,3642

R² = 0,1189

R² = 0,0003

56 57 58 59 60 61 62 63 64

0 0,05 0,1 0,15 0,2 0,25 0,3

HbA1c (mmol/mol)

Proportion prescription of the drug (%)

DPP-4 inhibitors (A10BH) GLP analogues (A10BJ) SGLT2- inhibitors (A10BK)

(33)

25 Figure 8. Correlation of HbA1c and proportion DPP4-inhibitors, GLP analogues and SGLT2- inhibitor prescribed during 2020 H1: The correlation of HbA1c and proportion of DPP4-inhibitors (A10BH), GLP analogues (A10BJ) and SGLT2-inhibitors (A10BK) prescribed from a PHC in Region Uppsala during 2020 H1. H1 stands for period January to June.

During 2020 H1, the correlation between HbA1c and the proportion DPP-4 inhibitors prescribed was R²=0,0042 (Figure 8). The correlation of GLP analogues prescribed and HbA1c was R²= -0,0976 and that of proportion SGLT2-inhibitors prescribed and HbA1c, R²=0,1431.

R² = 0,0042

R² = 0,0976

R² = 0,1431

52 54 56 58 60 62 64 66

0 0,1 0,2 0,3 0,4

HbA1c (mmol/mol)

DPP-4 inhibitors (A10BH) GLP analogues (A10BJ) SGLT2- inhibitors (A10BK) Linjär (DPP-4 inhibitors (A10BH))

(34)

26

4.7 Surveys

There was a low response on the surveys sent to nurses and physician that works at a PHC in Region Uppsala. A total of 13 out of 30 nurses that were sent the survey respond to the survey. Only 2 physicians respond the survey. This was probably due to the increased workload at the PHCs for these professions caused by the pandemic of Covid19.

4.7.1 Summary of survey sent to nurses at a PHC in Region Uppsala

Not all PHC in Region Uppsala have nurses specialized in diabetes care. The average number of nurses with such a position is one nurse at each PHC. Some PHC have 2-3 nurses,

although there are not many.

Out of the 13 nurses that respond to the survey, seven out of them respond that they are specialized in diabetes care. Two out of the seven comments that they are specializing in diabetes care. Among the six remaining nurses that respond to the survey, two of them have 60 credits of diabetes care while two others are district nurses. Overall, the nurses that respond to the survey sent to them have different amount of work experience in diabetes care that differs from less than 1-year experience to more than 10-year experience. Although four out of the 13 nurses that respond to the survey have 2 to 5-years of work experience in diabetes care.

Eight out of the 13 nurses that respond to the survey spends more than half of their worktime on diabetes care. Only two out of the 13 nurses did not mention diabetes rondes as a way of working together with physician specialized in diabetes care. 10 out of the 13 nurses work together with other nurses at the PHC, physician specialized in diabetes care and other physicians at the PHC. Three out of the 13 nurses respond that they work together with pharmacists. The same number of nurses works with medical technical companies. Seven out of 13 nurses work with medical companies. None of the nurses that respond to the survey works with authorities.

In general, the nurses that responded to the survey are content with the current work set-up.

Three out of the 13 nurses appreciate the diabetes rondes. One of the nurses likes the communication with the physicians. Another one responds that the collaboration with physician works well. One nurse appreciates the clear guidelines for treatment of diabetes.

Regarding things worth improving, three out of the 13 nurses responds that physicians should

(35)

27 be more updated of the new drugs and medical guidelines for treatment of diabetes. One nurse responds that physicians should prescribe Metformin more often. Another nurse responds that nurses specialized in diabetes care should be able to change drug dosage. One of the 13 nurses would like more time to go on training.

4.7.2 Summary of survey sent to physician at a PHC in Region Uppsala

Both physicians that respond to the survey are not specialized in diabetes care but they both have a degree in general medicine. One of the physicians also has a degree in intern

medicine. Both of physicians have 3 to 5-years of work experience with diabetes care. They respond that they do not spending much of their worktime on diabetes care. One of the physicians spends 5 % of their worktime on diabetes care while the other one spends an hour a week (did not respond in %). Physicians that respond to the survey forwarded to them works together with the same health personnel as the nurses. None of the two physicians that respond to the survey works with other physicians specialized in diabetes care, pharmacists, medical technical companies, or authorities.

The two physicians that respond to the survey are content with the current work set-up.The physicians are positive about all the drugs for treatment of T2DM. One of the physician’s states that SGLT2-inhibitors and GLP analogues have improved drug treatment of T2DM.

The physician’s states that more work focused on improved drug compliance and physicians and nurses receiving more training could improve diabetes care in Region Uppsala.

(36)

28

5.Discussion

5.1 Summary of main results

There is a difference in the prescribing patterns of diabetes drugs between different PHCs in Region Uppsala. This might be due to the strong association of the number prescribed prescriptions with the number of individuals at a PHC with a prescription, in this case a prescription for a drug with the ATC code A10. Although the difference in the prescribing patterns can be due to other factors as mention in the background. Individuals were

prescribed an average of two prescriptions of a drug with the ATC code A10 at a PHC in Region Uppsala during January 2018-June 2020. The average age of these individuals was 66 years and the average HbA1c-levels was 60 mmol/mol.

A larger proportion of OAD were prescribed compared to Insulins during the period January 2018- June 2020. Long acting insulin were prescribed to a greater extent compared to other Insulins at most PHC in Region Uppsala during 2018 H1 and 2020 H1. Between these periods, a larger proportion of Biguanide derivatives were prescribed compared to other OAD. During 2018 H1, almost the same proportion DPP-4 inhibitors were prescribed as SUs with some differences between PHC. While in 2020 H1, the proportion SGLT2-inhibitors prescribed were almost the same as that of DPP-4 inhibitors.

There were marginal differences in the proportion DPP-4 inhibitors, GLP analogues and SGLT-inhibitors prescribed dependent on the size, location, or income of the PHC. There was a low correlation between HbA1c and proportion DPP-4 inhibitors, GLP analogues and SGLT2-inhibitors prescribed at a PHC in Region Uppsala during 2018 H1 and 2020 H1.

What can be concluded from the surveys sent to the 13 nurses is that they are well qualified for their positions. They spend more than half of their worktime on diabetes care and works together with e.g. physicians by going on diabetes rondes. In general, the nurses that respond to the survey are content with their work and appreciates diabetes rondes. To mention

something worth improving 3 of the nurses’ responds that physicians should be more updated of the new drugs and medical guidelines for treatment of diabetes mellitus.

(37)

29 5.2 Discussion of the results

5.2.1 PHC characteristics

The number prescribed prescriptions differ from one PHC to another (Table 1). This results were expected as the number of prescribed prescriptions depends on the number of

individuals with a prescription for that drug at a PHC e.g. PHC P with 124 individuals PTL that prescribed a total of 423 prescription PTL compared to PHC G with 26 individuals PTL that prescribed 120 prescriptions PTL (Table 1). The difference in prescribed prescription might be due to other internal differences at the PHC e.g. the prescribing physician’s persistence to the drug and/or the difference in patients that needs the drugs (19).

Individuals were prescribed an average of two prescriptions of a drug with the ATC code A10 at a PHC in Region Uppsala during January 2018-June 2020. This is an appropriate number of prescriptions for treatment of T2DM as most diabetes medication should be used is combination to reach controlled glucose levels (13). Even though these results are positive, studies prove that individuals with T2DM receive more prescriptions than other individuals (25). Many individuals with T2DM have comorbid conditions, e.g. hypertension, which require other medical treatment. These increase the total number of prescribed prescriptions that an individual with T2DM receives, which might affect drug adherence (25). Extensive drug prescriptions can also be a financial burden, even when some expenses are covered by e.g. the government. This should be put into consideration when prescribing so that

individuals with T2DM can reach better glucose control (25).

The average age of individuals that were prescribed a prescription of a drug with the ATC code A10 in Region Uppsala between January 2018-June 2020 was 66 years (Table 1). These results were anticipated because increased age is one of the risk factors for T2DM (1). Where the risk of developing T2DM increase considerably after 45 years which might be due to the lack of physical activity by older people (26).

The average HbA1c-levels of individuals that were prescribed a prescription of a drug with the ATC code A10 in Region Uppsala between January 2018-June 2020 was 60 mmol/mol (Table 1) This result was anticipated for several reasons. One of those reasons is that HbA1c presented is an average of all HbA1c measures of individuals with a prescription of a drug with the ATC code A10 at a PHC in Region Uppsala between January 2018 – June 2020.

(38)

30 When the average HbA1c was calculated, there were no consideration made that the average HbA1c might be affected by other factors e.g. that some individuals might be newly

diagnosed with T2DM and therefor have a higher HbA1c than those who have been on treatment for a certain time. Apart from the reason mentioned, Socialstyrelsen states that half of the individuals with T2DM do not reach HbA1c reduction goals (18). This might be caused by unregular visits and lack of communication between patients with T2DM and healthcare. Therefore, increased visits and measures of HbA1c could be a way of improving that more patients reach HbA1c reduction goals. This includes offering patient training so that more patients learn how to live a healthier lifestyles and hopefully get better HbA1c reduction and employing more nurses and physicians specialized in diabetes care that can help treat these patients (18).

5.2.2 Proportion of Insulin (A10A) and OAD (A10B) prescribed

A larger proportion OAD were prescribed than Insulins during the period January 2018- June 2020 (Figure 2 compared to figure 3). The trend of proportion Insulins prescribed have decreased during this period while that of the proportion OAD prescribed have increased.

These results were anticipated as the first-line drug, Metformin, and most drugs in the recommendation list for treatment of T2DM are OAD and belongs to the ATC group A10B (8). Insulins are prescribed to a less extent for treatment of T2DM because these drugs are used as a supplement to help patients keep optimal glucose control (27). Patients with T2DM should be prescribed OAD with an intermediate- acting insulin, administrated before sleep to minimize the risk of hypoglycaemia (18).

5.2.3 Proportion of drugs with the ATC code A10A prescribed

A large proportion of long acting insulin compared to other Insulins were prescribed at most PHC in Region Uppsala during 2018 H1 and 2020 H2 (See appendix 3). These results were not anticipated as intermediate- acting insulin is the first line Insulin that should be prescribed to individuals with T2DM. Long acting insulins should be prescribed as a second resource when treatment with intermediate acting insulin is insufficient and patient still experience hyperglycaemia (18).

This result suggests that prescribing of long acting insulins at PHCs in Region Uppsala is not according to the recommendation list. The reason for this might be that these drugs are more favoured by physicians and patients compared to other Insulins leading to inappropriate

(39)

31 prescribing. Another reason for extent prescribing of long acting insulins in Region Uppsala might be that many people with a prescription for a drug with the ATC code A10 in the region might have more complicated diabetes and needs treatment with these drugs.

5.2.4 Proportion of drugs with the ATC code A10B prescribed

A large proportion of Biguanide derivatives were prescribed compared to other OAD during 2018 H1 and 2020 H2 (See appendix 4). These results were anticipated because the first-line drug in the recommendation list for treatment of T2DM, Metformin, belongs to the drug group Biguanide derivatives. This drug has a good safety profile and is cost-effective. It is the first-line drug that should be prescribed to all individuals with T2DM unless if they are elderly and fragile (8).

During 2018 H1, almost the same proportion DPP-4 inhibitors were prescribed as SUs with some differences between PHC. While in 2020 H1, the proportion of SGLT2-inhibitors prescribed were almost the same as the proportion DPP-4 inhibitors prescribed.

It was not anticipated that DPP4-inhibitors would be prescribed to a great extent in both 2018 H1 and 2020 H1 because these drugs have not been on the pharmaceutical market for long (13). This suggest that DPP4-inhibitors have been accepted for use by physicians and patients during a rather short period of time. Another reason for the extent prescribing of DPP4- inhibitors might be because they are the most cost-effective OADs that achieves good glycaemic control and weight reduction without causing hypoglycaemia (28).

More prescribing of SGLT2-inhibitors compared to SUs in 2020 H1 might be due to SGLT2- inhibitors positive pharmacological effects. These drugs achieves good glycaemic control and weight reduction without causing hypoglycaemia, just as DPP4-inhibitors (28). SGLT2- inhibitors have positive effects on the CV which suggests that these drugs might soon be prescribed for other indication e.g. chronic heart failure (30). SUs are more cost-effective than SGLT2-inhibitors, but they cause hypoglycemia and weight gain (14) which might be the reason to lower prescribing of these drugs during 2020 H1.

(40)

32 5.2.5 Difference in prescribing of DPP4-inhibitors, GLP analogues and SGLT2-

inhibitors

There were marginal differences in the proportion DPP-4 inhibitors, GLP analogues and SGLT2-inhibitors prescribed at small, middle and big sized PHC in Region Uppsala during 2020 H1 (Figure 4). The same differences were observed between PHC in the countryside and those in urban areas in the region during this period (Figure 5). Suggesting that the prescribing pattern of diabetes medication in Region Uppsala during 2020 H1 is not much affected by the size of the PHC or were the PHC is situated in this study. The reason for the marginal difference in the proportion DPP-4 inhibitors, GLP analogues and SGLT-inhibitors prescribed depending on the two variables might be due to different medical conditions among patients at the PHC. Because these drugs are often prescribed to individuals with other medical conditions as heart failure, obesity or kidney failure (8).

There was a slight difference in proportion DPP-4 inhibitors, GLP analogues and SGLT2- inhibitors prescribed based on the income before tax of the municipality where the individual received the prescription from, during 2020 H1 (Figure 6). This result suggest that income is not an important factor that affects prescribing of diabetes medication in Region Uppsala during this period.

Although income does not seem to affect prescribing in this study, income affects what type of drugs an individual can afford. Even in a country as Sweden were the government pays for a portion of the drug through “högkostandsskyddet” (31), medical treatment for T2DM can still be expensive for those with a low income. Especially when individuals with T2DM often have comorbid conditions e.g. hypertension (25). It is therefore important that the prescriber has in mind the financial ability of the patient when prescribing a certain drug.

5.2.6 Correlation of HbA1c with proportion of DPP-4 inhibitors, GLP analogues and SGLT2-inhibitors prescribed

The correlations between HbA1c and proportion of DPP-4 inhibitors, GLP analogues and SGLT2-inhibitors that were prescribed at a PHC in Region Uppsala was low for all the drug groups, during 2018 H1and 2020 H1 (See figure 7&8). The association between HbA1c and proportion of drug prescribed was expected to be generally low because there are several factors that affects HbA1c than drug prescribed such as the patients general health (8).

(41)

33 5.2.7 Survey sent to nurses and physicians

The results from the survey sent to the nurses proves that the 13 nurses that respond to the survey are well qualified for their positions and spend more than half of their worktime on diabetes care. Even though these results are positive, Nationella riktlinjer för diabetesvård states that more nurses specialized in diabetes care are needed so that HbA1c-reduction goals can be met at PHC (18). The same cannot be conclude from the survey sent to the physicians, due to low response rate.

From the response of the surveys sent to nurses and physicians one can conclude that there is not enough collaboration between these health personals and pharmacists at PHCs in Region Uppsala. This is something that could be improved as pharmacists have proven to positively affect drug prescribing at health practices (32).

5.3 Strength, weaknesses, and limitations

5.3.1 Strengths

A cross sectional study was conducted to study whether there is a prescribing pattern of diabetes drugs at PHC in Region Uppsala. This was the most appropriate method to use for a study that have never been conducted before on the population in Region Uppsala to identify possible factors that might affect the prescribing patterns of T2DMD. The negative side of conducting a cross sectional study is that the association found can be casual or not.

Therefore, further investigation should be conducted to be able to determine the causality of the association. This can be done by conducting analytical studies e.g. random controlled trails (RCT), which is an expensive method that would not have been feasible to conduct for this project.

It was a strength to have Region Uppsala’s electronic health records (EHR) as a source for data. Even though this EHR have never been used for such a purpose before, it is of great value that the region has access to such data. Another strength with this study is the big population size which indicates that the results can be generalized for Region Uppsala.

Access to data for the five periods have proven to be a strength of this study. Some analysis was conducted during different periods which made it possible to see changes that occurred over time.