Sessio 9

Sessio 9: Electronic patient records and national registers in assessment and development of quality of care

Puheenjohtaja: professori Tiina Laatikainen, Itä-Suomen yliopisto

9.1 LDL cholesterol trajectories and statin use in Finnish type 2 diabetes patients – a longitudinal cluster analysis

Laura Inglin, Piia Lavikainen, Kari Jalkanen, Marja-Leena Lamidi, Tiina Laatikainen

Aims:
This study aimed to identify groups of type 2 diabetes patients with similar longitudinal trends in low-density lipoprotein cholesterol (LDL-C) levels and investigate the association of these trajectories with statin therapy and treatment changes.

Methods:
This retrospective cohort study utilized electronic health records (EHRs) from patients with type 2 diabetes in North Karelia, Finland, comprising all primary and specialized care visits during 2011-2017. LDL-C measurements and statin prescriptions were extracted from the EHRs during 2012-2017. Latent class mixed model was applied to identify distinct LDL-C trajectories.

Results:
From the cohort, 4652 men and 3991 women with at least two LDL-C measurements during the follow-up were identified, and included in the study. Four LDL-C trajectories were extracted: “moderate-stable”, “decreasing”, “high-stable”, and “increasing”. Over 80% of the patients had “moderate-stable” LDL-C levels around 2.3 mmol/L during the follow-up, with high overall prevalence of moderate- and high-intensity statin treatment. In 5% of patients, a steep “decrease” in initially very high LDL-C levels was observed, along with a steady increase in moderate- and high-intensity statin treatment. Patients in the “high-stable” group with an average LDL-C level of 3.9 mmol/L over the follow-up had the lowest proportion of yearly measured LDL-C, and the highest proportion of patients without statin treatment. Patients in the “increasing” LDL-C trajectory had the highest rates of treatment de-intensification during the follow-up. The two groups with poor or deteriorating LDL-C control were significantly more frequent among women than men. Overall as well as in most trajectories, women had significantly higher average LDL-C levels, more frequently no statin treatment, and less frequently high-intensity treatment than men.

Conclusion:
The majority of patients had LDL-C controlled with improving tendency. More attention should be paid to patients with high-stable LDL-C values, especially women, by providing motivating counselling and by intensifying their lipid-lowering medication.

9.2 Treatment lines and the quality of type 2 diabetes care among patients with multimorbidity

Nazma Akter Nazu, Katja Wikström, Marja-Leena Lamidi, Jaana Lindström, Hilkka Tirkkonen, Päivi Rautiainen, Tiina Laatikainen

Aims:
The study assessed the effect of different recommended treatment lines on quality of diabetes care among the primary care cohort of type 2 diabetes (T2D) patients with and without multimorbidity in North Karelia, Finland, from 2011-12 to 2015-16.

Methods:
Data on all diagnosed T2D patients (n=10190) were collated retrospectively from regional electronic health records (EHR). Information on prescribed treatment was received from data on e-prescriptions. The used treatments, their effects and possible intensification of treatments (hypoglycaemic agents and lipid-lowering agents) on patients with only T2D, with any mental disorder, with cardiovascular diseases (CVD), and with CVD and any mental disorder were analysed.

Results:
The proportion of patients without hypoglycaemic agent decreased during the follow-up except among those having T2D with CVD and mental disorder. Metformin was the most used medication for all of the patient groups except the T2D patients with CVD and mental disorder. The use of metformin decreased as the disease progressed in all patient groups. T2D patients with only CVD and combined CVD and mental disorders were mostly using a combination of tablets and insulin and the use increased during follow-up. A considerable number of patients were not using any lipid-lowering agent. For lipid lowering the moderate intensity agents were mostly used irrespective of the patient groups. The use of high-intensity lipid-lowering agents was observed only among T2D patients with CVD.

Conclusions:
Glucose management treatments need to be intensified as the disease progress. However, there are a considerable number of patients who are not using any medication most probably because they have mild diabetes and good glycaemic control. Patients with CVD are mostly treated with more intensified treatment, but patients with CVD and mental disorder are mostly without medication indicating a treatment gap. The risk of CVD should be better considered also among these patients.

Keywords: quality of care, mental disorder, type 2 diabetes, cardiovascular disease, mode of treatment, hypoglycaemic agent, lipid-lowering agent.

9.3 The impact of varying care practices for ST-elevation myocardial infarction (STEMI) on socioeconomic equity in outcomes of care

Eeva Reissell, Sonja Lumme, Kristiina Manderbacka

Background
Previous studies have shown persistent regional variation and socioeconomic differences in the use of revascularisations for ST-elevation myocardial infarction (STEMI). However, studies on regional equity in percutaneous coronary interventions (PCI) taking supply side factors and the need for care into account are scarce. In this study we aim to investigate whether the impact of varying care practices for STEMI or regional demand including socioeconomic factors play a larger role in patients’ short- and long-term mortality.

Methods
Individual-level data of patients aged > 40 admitted to a hospital due to STEMI in 2015–17 in Finland were obtained from the Finnish Care Register for Health Care and linked to register data on socio-demographics from Statistics Finland. 30-day and 365-day case-fatality after STEMI are used as indicators of quality of care. The data is analysed using survival analysis, Poisson regression and multilevel modelling. Additionally area-level factors are used to study care practices in hospital districts.

Results
We found clear socioeconomic differences in 30- and 365-day mortality after STEMI, which remained after controlling for care practices for STEMI among men. The risk for CHD death increased gradually from highest to lowest income group. RR for men was 2.36 (95% CI 1.57-3.52) for lowest income quintile compared with the highest quintile for 365-day mortality. The 30-day mortality for men in the lowest income group was 2.06 (1.37-3.08) compared to the highest. No socioeconomic differences between income groups were found among women. Regional differences in mortality were minor.

Discussion
As in previous studies, we found regional variation in the use of PCI for STEMI. We detected systematic socioeconomic differences both in 30-day and 1 year mortality after STEMI in men but not in women. The observed regional differences in care practices did not explain the effect of socioeconomic status on case fatality in men.

9.4 Development of national quality register for diabetes care

Aapo Tahkola, Sari Koski, Saara Metso, Seija Olli, Tiina Laatikainen

The development of national healthcare quality registers is of high importance in Finland. However, Finland currently lacks tools for system- and nationwide healthcare quality monitoring. In 2018, Finnish parliament granted special funding for the National Institute for Health and Welfare (THL) to promote the development of healthcare quality registers. The National healthcare quality registers –project was launched at the same year. It aims to create a national model for permanent quality registers via seven pilot register projects, one of which concerns diabetes care.

Diabetes quality register pilot -group started working in January 2019 by studying the existing international examples of diabetes quality registers and on-going quality improvement work in Finnish health care organizations. Register inclusion criteria was defined and tested in practice with real world register data from three different hospital districts and one nationally operating private health care organization. An agreement on the inclusion criteria, minimum data content, a preliminary plan for the diabetes quality indicators and testing the indicators with real world data were performed. The results of the first analysis concerning diabetes type 2 were presented on a quality improvement meeting in April 2020 gathering a wide range of representatives from all levels of health care system and from 13 municipalities.

Preliminary results show that compared with other countries both LDL and glucose control measured by HbA1c are quite good in Finland. However, there is notable variation between municipalities and especially the management of LDL needs improvement.

It seems possible and feasible to build national diabetes quality register on the basis of routinely recorded clinical data. However, a lot of work and actual resources combined with successful co-operation between register maintenance, scientific research and practical quality improvement are still needed to establish a high-functioning and permanent national quality register for improving diabetes care.

9.5 Trends in regional variation in elective hip and knee arthroplasties 2010-2017 in Finland

Kristiina Manderbacka, Martti Arffman, Markku Satokangas, Eeva Reissell, Ilmo Keskimäki

Background:
A persistent research finding in Finland and elsewhere has been regional variation in medical practices. We examined over time variation in elective primary hip (THA) and knee arthroplasties (TKA) and possible explanatory factors related to individual comorbidities and socioeconomics (SEP) as well as area-level disease burden, spatial access and supply of hospital care.

Methods:
Data of patients undergoing primary THA and TKA among the total Finnish population aged >25 in 2010-2017 was obtained from the Finnish Care Register for Health Care and Finnish Arthroplasty Register. This data was linked with individual SEP – obtained from FOLK modules of Statistics Finland – and area-level factors calculated for hospital districts. Using multilevel Poisson regression models we analysed the proportion of variance explained by individual and area-level factors separately in 2010-2013 and 2014-2017.

Results:
In gender- and age-adjusted THA models in 2010-2013, individual SEP explained 15% of the hospital district variance, area-level proportion of 65+olds further 17% and musculoskeletal disorder index (MDI) further 32%. In 2014-2017, individual factors did not explain any variance, while proportion of persons aged 65+ explained 8%, MDI further 5% and having university hospital in the district further 13%. In TKA the variance was larger. In 2010-2013, individual SEP (12%), proportion of 65+olds (15%), MDI (18%) and number of orthopaedic surgeons in the district (10%) were important explanatory factors. In 2014-2017, individual SEP (6%), proportion of 65+olds (10%), MDI (18%) and having university hospital in the district (16%) were prominent.

Discussion:
Alongside individual socioeconomic position, area-level disease burden and factors related to supply of services seem to play a role in regional variations in access to elective THA and TKA procedures. While factors examined in this study explained 78% of differences in THA in 2010-2013, a large part of the variance in TKA and in THA 2014-2017 remained unexplained.