Inger Johanne Bakken

Background: Type 1 diabetes (T1D) and type 2 diabetes (T2D) differ in health care needs. Precise prevalence estimates by type are important for health care planning. Data and methods: This validation study used Norwegian health registry data to develop a T1D/T2D classification method, using the Norwegian Diabetes Register for Adults (NDR-A) as reference standard. Our study population consisted of all individuals 18 years and older residing in Norway in 2022 ( N = 4,315,563). We used Cohen’s kappa to assess agreement between the classification method and the NDR-A. Results: The established classification method had three key steps: (a) T1D, based on insulin use and specialist health care T1D diagnoses, (b) T2D, based on specialist and primary health care T2D, and (c) T1D, based on primary health care T1D and insulin use, and no T2D. The overall agreement with NDR-A diagnosis was 95.2% for T1D and 97.6% for T2D. For T1D, agreement was 97% or higher in age-groups 18–39 and 40–59, 87.6% for those aged 60–79 and 69.4% for individuals older than 80. For T2D, the agreement was 97% or higher for age-groups 40 and above, and 92.1% for those aged 18–39. Applying the classification method to total population data showed that, in 2022, among 4,315,563 individuals aged 18 and older living in Norway, 31,264 had T1D (0.72%) and 239,120 had T2D (5.5%). Conclusions: Our registry-based classification agreed well with the NDR-A, implying that administrative health registry data are suitable in epidemiological and health services studies of T1D and T2D.

Aim: Timely monitoring of maternal vaccination is essential for guiding immunization strategies, but ongoing pregnancies are usually not identified in health registries, resulting in delayed surveillance. We evaluated a registry-based method using routinely collected pregnancy-specific ultrasound procedure codes introduced in 2022 to identify ongoing pregnancies and demonstrate feasibility of timely vaccination surveillance, using influenza vaccination as an example. Methods: We linked population-based registry data from the Medical Birth Registry of Norway (MBRN), the Norwegian Patient Register (NPR), and the Immunisation Registry. Ongoing pregnancies in the NPR were identified using either second-trimester ultrasound procedure codes (Scenario 1) or a combination of first- and second-trimester codes with adjustments for pregnancy loss and non-attendance (Scenario 2). Estimates of vaccination uptake using ongoing pregnancies identified in the NPR as denominator were compared with estimates using birth data from the MBRN data as denominator (gold standard). Monthly and cumulative seasonal influenza vaccination uptake was assessed from October 2022 to February 2023. Results: A total of 26,996 and 27,940 pregnant women were identified using Scenarios 1 and 2, respectively, compared with 27,513 using MBRN data. Monthly vaccination uptake estimates deviated by less than 1.5% from the gold standard. Cumulative vaccination uptake was 27.1% (7328/26,996) and 26.9% (7503/27,940) for Scenarios 1 and 2, compared with 27.8% (7656/27,513) using MBRN data. Conclusions: Ongoing pregnancies can be identified through ultrasound procedure codes from administrative healthcare data, enabling robust and timely vaccination surveillance during pregnancy. The approach supports near real-time vaccination monitoring and might be applicable to broader antenatal care surveillance in general.

Background: Bariatric surgery provides notable weight loss and metabolic benefits but may also increase the risk of alcohol use disorder (AUD). We investigated whether Roux-en-Y gastric bypass (RYGB) confers higher rates of AUD and related disorders than sleeve gastrectomy (SG) and examined associated morbidity and mortality. Methods: We conducted a retrospective, population-based cohort study using data from the Norwegian Patient Registry and the Norwegian Prescription Database. A total of 17,800 patients underwent RYGB ( n = 12,244) or SG ( n = 5556) between 2008 and 2018. Patients with prior diagnoses related to alcohol or use of medications for AUD were excluded. Incidence rates (IR) for alcohol-related diagnoses were calculated per 1000 person-years; hazard ratios (HR) were derived comparing RYGB to SG. Morbidity was measured as number of specialized healthcare contacts, but also as number of prescriptions/defined daily doses in an exploratory model. Results: Mean postoperative follow-up was 5.7 years (RYGB) and 3.8 years (SG). By 31 December 2018, 576 patients (3.2%) had developed a new alcohol-related diagnosis – an incidence rate of 6.34 per 1000 person-years. The adjusted HR for such diagnoses was 1.69 (95% CI 1.33–2.13, p < 0.001) for patients undergoing RYGB compared to SG. Because mortality did not differ significantly between RYGB and SG, mortality was assessed for the cohort as a whole: patients with alcohol-related diagnoses had an adjusted HR for death of 2.08 (95% CI 1.40–3.08) relative to those without. They also recorded, on average, 5.5 additional contacts in specialist care. Conclusions: Compared with SG, RYGB was associated with a 69% higher risk of alcohol-related diagnoses. Further, given the elevated morbidity and mortality linked to these disorders, enhanced preoperative screening and long-term postoperative monitoring are warranted in modern bariatric practice.

Objectives: To investigate the clinical characteristics of weight loss drugs (WLDs) users in Norway. Design: Nested population‐based case–control study. Setting: Nationwide healthcare registries with data on dispensed medications and healthcare diagnoses. Participants: Individuals aged 18–74 years initiating WLDs during 2023–2024, categorised by drugs: semaglutide (only Wegovy), liraglutide (only Saxenda), tirzepatide, bupropion‐naltrexone or orlistat. Each WLD user was matched with five age‐ and sex‐matched population controls. Main Outcome Measures: Comorbidities diagnosed in the 2 years preceding initiation of WLD use and comedications dispensed in the preceding year, analysed using conditional logistic regression. Results: 150 036 individuals initiated semaglutide, 4603 liraglutide, 3596 tirzepatide, 31 172 bupropion–naltrexone and 1411 orlistat. Among WLD users, 71%–78% were registered with at least one comorbidity, compared with 49%–51% in controls. Nineteen percent to 24% of WLD users had four or more comorbidities. Hypertension was observed among 32%–38% of WLD users versus 17%–20% of controls; hyperlipidaemias: 20%–24% versus 12%–15%; sleep apnoea: 7%–11% versus 2%–3%; back pain: 15%–20% versus 10%–11%; opioids: 24%–30% versus 13%–15%; antidepressants: 20%–24% versus 10%–11%. Having only a basic level of education (≤ 10 years) was more common among WLD users than among controls (20%–28% vs. 17%), except for tirzepatide users, where the education level distribution was as for controls. Conclusions: WLD use is rapidly increasing, with users often having complex comorbidities. Understanding these patterns helps ensure that medications are used as intended. Like other countries with universal health systems, Norway must address the high cost of obesity treatments and mitigate health disparities to ensure access for those most likely to benefit.

Aims: To describe updated trends in the incidence and prevalence of diagnosed type 2 diabetes in Norway by age, sex, country of birth and education. Methods: A nationwide cohort study using registries on primary and specialist healthcare, dispensed drugs and demographic factors in Norway. We analysed incidence trends 2009–2021 using Poisson regression and JoinPoint. Results: During 2009–2021, 195,935 incident type 2 diabetes cases were identified. After a decline in incidence 2009–2014, the incidence was largely stable over time, with a suggestive upward trend 2019–2021. The overall incidence rates were 651 per 100,000 person‐years in 2009 and 530 per 100,000 in 2021. There were 165,432 prevalent cases of type 2 diabetes in 2009 (5.5% of the population), consistently increasing throughout the study period to a peak at 259017 (7.6%) in 2021. The time trends were largely consistent across age, sex, country of birth and education. Inhabitants with lower education and born in Asia or Africa had substantially higher incidence and prevalence than those with higher education and born in Norway or other continents, respectively. Conclusions: The previously described decline in incidence of diagnosed type 2 diabetes during 2009–2014 was not sustained, and prevalence continued to increase throughout 2009–2021.

Objectives: To investigate clinical characteristics of users of weight loss drugs in Norway. Design: Nested population-based case control study Setting: Nationwide healthcare registers in Norway with information on all dispensed medications linked to contacts with primary and specialist healthcare services. Participants: All individuals aged 18-74 years with first dispensing of a weight loss drug (WLD) in 2023–2024, classified as initiators of 1) semaglutide (only Wegovy), 2) liraglutide (only Saxenda), 3) tirzepatide, 4) bupropion–naltrexone, or 5) orlistat. We matched each WLD user by sex and age to five controls randomly selected from the general Norwegian population. Main outcome measures: Type and count of number of comorbidities diagnosed in the two years preceding initiation of WLD use, comedications dispensed in the preceding year, and any previous bariatric surgery. Differences between WLD users and population controls were estimated using conditional logistic regression adjusted for country of birth and education level. Results: During 2023-24, 150,036 individuals initiated semaglutide, 4,603 liraglutide, 3,596 tirzepatide, 31,172 bupropion–naltrexone, and 1,411 orlistat. Among WLD users, 22-29% were registered without any of the pre-defined comorbidities, compared to around half of population controls. Comorbidities and comedications were generally observed at similar proportions in the different WLD user groups, but at much lower proportion for controls: hypertension: 32-38% of WLD users vs. 17-20% of controls; hyperlipidaemias: 20-24% vs. 12-15%; sleep apnoea: 7-11% vs. 2-3%; back pain: 15-20% vs. 10-11%; opioids: 24-30%vs. 13-15%; antidepressants: 20-24%vs. 10-11%. Conclusions: Like other countries, Norway faces the challenge of costly pharmacological obesity treatment, and the need for measures to mitigate increasing socioeconomic health disparities to ensure access to obesity treatment for those most likely to benefit including persons with multimorbidity. What is already known on this topic: Obesity is increasing worldwide and is associated with increased risk of multimorbidity. The available treatment options for obesity have been rapidly and substantially transformed by the availability of new weight loss drugs. What this study adds: This nested case-control study within the Norwegian population examined clinical characteristics of new users of five weight loss drugs: semaglutide, liraglutide, tirzepatide, bupropion–naltrexone, and orlistat. The study identified a higher proportion of weight loss drugs users with cardiovascular, endocrine, mental, and other health conditions, as well as increased usage of cardiovascular, nervous system, and chronic pain medications, compared to the general population. The majority of WLD users had high levels of multimorbidity and medication use compared to controls. Approximately 1 in 5 WLD users had no registered comorbidity in either primary or specialist health care systems. Additionally, we noted differences in sociodemographic factors between users and population controls, such as lower education levels among WLD users. The rapidly expanding use of weight loss drugs necessitates close attention from public health authorities.

Background: Type 1 diabetes (T1D) and type 2 diabetes (T2D) differ in health care needs, and precise prevalence estimates by diabetes type are important for health care planning. Data and Methods: Using Norwegian health registry data on diagnoses from primary and specialist care and prescribed drugs, we developed a classification method to define T1D and T2D. The Norwegian Diabetes Register for Adults (NDR-A) served as reference standard. Results: First, for individuals in the NDR-A, we established a health registry data classification method with three key steps: (A) classifying T1D based on the use of insulin and diagnoses from specialist healthcare, (B) classifying T2D based on diagnoses from both specialist and primary healthcare, and (C) identifying individuals as having T1D if they were diagnosed in primary healthcare and used insulin, but never had a T2D diagnosis. Overall correspondence between NDR-A diagnosis and the classification method was 94.7% for T1D and 97.9% for T2D. For T1D, the correspondence was 97% or higher in age-groups 18-39 and 40-59, 87.6% for those aged 60-79 and 69.4% for individuals older than 80 years. For T2D, the correspondence was 97% or higher for age-groups 40 and above, and 92.1% for those aged 18-39. Applying the classification method to total population data showed that, in 2022, among 4,315,563 individuals aged 18 years and older living in Norway, 31,264 were identified as having T1D (0.72%) and 239,120 were identified as havingT2D (5.5%). Conclusion: While classifying diabetes types can be challenging, our registry-based classification corresponded well with the NDR-A.

Aims: To identify an efficient file format for data delivery from large administrative registers for research, testing the full workflow from data extraction to research data management. Methods: Through collaboration between a data delivery and a research department within the same institute, we evaluated each step from data extraction and delivery to management and usage, comparing CSV and Parquet formats. Results: Switching from the unstructured, text-based CSV format to the highly structured Parquet format significantly optimized all processes by reducing file sizes and saving processing time. The Parquet format also provided access to advanced data management techniques, simplifying further work. Despite these advantages, the basic programming required for Parquet format is not very different from that for CSV. We provide a tutorial and examples as online supplement. Conclusions: We strongly recommend replacing CSV files with contemporary data formats. The Parquet file format proved to be an excellent option throughout the entire process from data extraction to research implementation

Aims/hypothesis: To investigate updated trends in the incidence and prevalence of diagnosed type 2 diabetes in Norway by age, sex, country of birth and education. Methods: A nationwide cohort study using registries on primary and specialist healthcare, dispensed drugs, and demographic factors in Norway. We analysed incidence trends 2009-2021 using Poisson regression and JoinPoint. Results: During 2009–2021, 195 935 incident type 2 diabetes cases were identified. After a decline in incidence 2009–2014, the incidence was largely stable over time, with a suggestive upward trend 2019–2021. The overall incidence rates were 651 per 100 000 person-years in 2009 and 530 per 100 000 in 2021. There were 165 432 prevalent cases of type 2 diabetes in 2009 (5.5% of the population), consistently increasing throughout the study period to a peak at 259 017 (7.5%) in 2021. The time trends were largely consistent across age, sex, country of birth and education. Inhabitants with lower education and born in Asia or Africa had substantially higher incidence and prevalence than those with higher education and born in Norway or other continents, respectively. Conclusions/interpretation: The previously described decline in incidence of diagnosed type 2 diabetes during 2009– 2014 was not sustained, and prevalence continued to increase throughout 2009–2021. Research in context: What is already known about this subject?: The incidence rate of type 2 diabetes in Norway and several other countries decreased from 2009–2014, despite an ongoing increase in prevalence. What is the key question?: Has the decrease in incidence rate of type 2 diabetes in Norway continued after 2014? What are the new findings?: The incidence rate of type 2 diabetes in 30–89-year-old residents in Norway was relatively stable from 2015–2021, with a suggestive increase from 2019–2021, largely consistent by sex, age, education and country of birth. The prevalence continued to increase. How might this impact on clinical practice in the foreseeable future?: Both incidence and prevalence are important for shorter term health care planning in an ageing population and to understand time trends.