Skip to main content
PLOS ONE logoLink to PLOS ONE
. 2022 Jun 1;17(6):e0269245. doi: 10.1371/journal.pone.0269245

Data-driven long-term glycaemic control trajectories and their associated health and economic outcomes in Finnish patients with incident type 2 diabetes

Piia Lavikainen 1,*, Emma Aarnio 1, Miika Linna 2, Kari Jalkanen 1, Hilkka Tirkkonen 3, Päivi Rautiainen 3, Tiina Laatikainen 3,4,5, Janne Martikainen 1
Editor: Dylan A Mordaunt6
PMCID: PMC9159579  PMID: 35648780

Abstract

Background

Treatments should be customized to patients to improve patients’ health outcomes and maximize the treatment benefits. We aimed to identify meaningful data-driven trajectories of incident type 2 diabetes patients with similarities in glycated haemoglobin (HbA1c) patterns since diagnosis and to examine their clinical and economic relevance.

Materials and methods

A cohort of 1540 patients diagnosed in 2011–2012 was retrieved from electronic health records covering primary and specialized healthcare in the North Karelia region, Finland. EHRs data were compiled with medication purchase data. Average HbA1c levels, use of medications, and incidence of micro- and macrovascular complications and deaths were measured annually for seven years since T2D diagnosis. Trajectories were identified applying latent class growth models. Differences in 4-year cumulative healthcare costs with 95% confidence intervals (CIs) were estimated with non-parametric bootstrapping.

Results

Four distinct trajectories of HbA1c development during 7 years after T2D diagnosis were extracted: patients with “Stable, adequate” (66.1%), “Slowly deteriorating” (24.3%), and “Rapidly deteriorating” glycaemic control (6.2%) as well as “Late diagnosed” patients (3.4%). During the same period, 2.2 (95% CI 1.9–2.6) deaths per 100 person-years occurred in the “Stable, adequate” trajectory increasing to 3.2 (2.4–4.0) in the “Slowly deteriorating”, 4.7 (3.1–6.9) in the “Rapidly deteriorating” and 5.2 (2.9–8.7) in the “Late diagnosed” trajectory. Similarly, 3.5 (95% CI 3.0–4.0) micro- and macrovascular complications per 100 person-years occurred in the “Stable, adequate” trajectory increasing to 5.1 (4.1–6.2) in the “Slowly deteriorating”, 5.5 (3.6–8.1) in the “Rapidly deteriorating” and 7.3 (4.3–11.8) in the “Late diagnosed” trajectory. Patients in the “Stable, adequate” trajectory had lower accumulated 4-year medication costs than other patients.

Conclusions

Data-driven patient trajectories have clinical and economic relevance and could be utilized as a step towards personalized medicine instead of the common “one-fits-for-all” treatment practices.

Introduction

Main target of type 2 diabetes (T2D) care is to achieve and maintain good treatment balance to avoid complications and premature deaths [14]. Typically, glycated haemoglobin A1c (HbA1c) is used as a measure of long-term blood sugar level with values <53 mmol/mol (7.0%) indicating good treatment balance. However, individualized, more-stringent goals for young T2D patients with short disease duration and less-rigorous, personalized treatment targets for older and the most comorbid patients are recommended [1, 4]. Treatments should be customized to patients to improve patients’ health outcomes and maximize the treatment benefits.

While the need of individual tailoring of treatments due to heterogeneity of T2D patients is recognized, majority of patients still receive standardized “one-size-fits-all” care due to limitations in available resources for individualized treatment. However, identification of patient groups with similarities in care needs and treatment outcomes, taking also into account budget constraints and cost consequences, is a step towards personalized care. It is also economically rational to identify subgroups where the treatment is cost-effective [5] and to offer care accordingly. Clustering of treatment outcomes, namely HbA1c, of T2D patients provides tools for treatment targeting and service design, and, further, optimizing the allocation of limited economic resources. Patients in different trajectories are likely to prefer different configurations of diabetes care and support. For example, patients with a deteriorating treatment balance profile should receive more intensive disease management supported by health professionals whereas adequate self-management support could be well enough for patients within stable treatment balance.

The objective of this study was to identify meaningful data-driven trajectories of incident T2D patients with similarities in HbA1c patterns since diagnosis. Further, we aimed to describe incidence of diabetes complications, all-cause mortality, and accumulation of healthcare costs by the trajectories to validate clinical applicability of the results in optimization and supporting effectiveness of treatment.

Materials and methods

Study setting

In Finland, most health services are public meaning all residents have the right to use and equal access to them. Municipalities are responsible for organizing and financing public health and social care services for their residents. Siun sote—Joint municipal authority for North Karelia social and health services was established in the region of North Karelia at the beginning of 2017 [6]. Instead of one municipality, Siun sote joins healthcare organisations of 13 municipalities with around 166 000 inhabitants. Before the integration to Siun sote, all the municipalities in the North Karelia region had introduced an identical electronic health register (EHR) called Mediatri in 2010–2011, and its use has been continued also after the integration. EHR data from the Mediatri cover both public primary and specialized healthcare. However, these EHRs do not cover patients utilizing only private healthcare services that produce about one quarter of all the health and social care services.

All community-dwelling residents of Finland are eligible for medication reimbursement, and reimbursed medication purchases from community pharmacies are registered in the Finnish Prescription Register (FPR) established in 1994 and maintained by the Social Insurance Institution of Finland (SII). However, patients staying at public nursing homes or hospitals >90 consecutive days are not eligible for medication reimbursement and their purchases are not registered during these stays. In addition, the FPR does not include records of over-the-counter medication purchases. The Special Reimbursement Register (SRR), established in 1964 and also maintained by the SII, includes records of entitlements to higher medication reimbursements due to certain chronic diseases, such as T2D.

Study design

This study utilized data provided by Siun sote, where the prevalence of T2D was 6.2% in 2012 [7]. All patients with newly diagnosed T2D in 2011–2012 were identified with ICD-10 code [8] E11 from the EHRs. Collected EHR data contained information on diagnoses, medication prescriptions, as well as laboratory assays during 2010–2017. EHR data were combined with information on medication purchases (such as dispensing date, Anatomical Therapeutic Chemical (ATC) classification code [9], and cost) from the FPR for the years 1995–2017. Entitlements to special refund for diabetes medications prior to 2011 were retrieved from the SRR.

Timing of T2D diagnosis was ascertained with data on diabetes medication purchases (ATC code A10) from the FPR in 1995–2010 and the retrieved entitlements to special refund for diabetes medications since 1964. If a patient had T2D diagnosed by healthcare providers outside of Siun sote, for example, before moving to the North Karelia region, the first record of T2D could be tracked from the national FPR and SRR instead of the first record in the EHRs of Siun sote. T2D diagnosis date was considered to be the first ever record of either diabetes medication purchase in the FPR, entitlement to special refund in the SRR, or T2D diagnosis in the EHRs, and only patients with an ascertained diagnosis recorded in 2011–2012 were included.

Glycaemic control

Glycaemic control was measured with HbA1c with the turbidimetric inhibition immunoassay method (TINIA) in the Eastern Finland laboratory (ISLAB, https://www.islab.fi). The method and its analysis were standardized across the North Karelia region. Values were standardized to the International Federation of Clinical Chemistry (IFCC) units. EHRs data on measurements from 3 months prior to T2D diagnosis until a maximum of 89 months after it were used. Patients with ≥1 HbA1c measurement during the study period were included. Baseline HbA1c (mmol/mol) level was defined to be the one measured ±3 months from T2D diagnosis. Follow-up measures were categorized as occurring 6–17 (1 year), 18–29 (2 years), 30–41 (3 years), 42–53 (4 years), 54–65 (5 years), 66–77 (6 years), and 78–89 months (7 years) after diagnosis. If several measurements occurred within a period, the latest one was selected. Sensitivity analyses were conducted against the average of the HbA1c measurements within a period and the results were comparable. If HbA1c value was unmeasured within a certain period, it was considered as a missing value. HbA1c values were measured until the date of death, moving outside of the study area, or 31st December 2017, whichever occurred first.

Deaths and diabetes complications

Information on the dates of deaths and complications occurring after T2D diagnosis were retrieved from the EHRs. Micro- and macrovascular complications were identified with ICD-10 codes which have been in use in Finland since 1996 (S1 Table). Complications were examined as a composite of micro- and macrovascular complications as well as separately. Incidence of all-cause deaths and diabetes complications were followed up until the date of the specific event, moving outside of the study area, or 31st December 2017, whichever occurred first.

Healthcare costs

Healthcare costs were calculated as a cumulative sum of primary and specialized healthcare costs, nursing home and home care costs, and outpatient medication costs during 2014–2017. EHR data included all hospitalizations (outpatient visits, inpatient admissions) and their costs derived from the hospitals’ cost accounting systems in 2014–2017. Restriction in cost analyses to years 2014–2017 was applied because data on outpatient visits were not complete and, therefore, comparable between municipalities during the transitional phase to the Mediatri before 2014. The individual-level resource use data for primary care were also obtained from the EHRs (with diagnosis and activity information) and grouped using the Ambulatory and Primary Care-Related Patient Groups (APR) grouper software, a grouping system equivalent to DRG used in hospital care. The unit cost estimates for grouped primary care contacts were obtained from the national standard price lists for primary care encounters [10]. Outpatient medication costs were calculated based on the information on the total costs of all reimbursed as well as reimbursed antidiabetic (ATC code A10) medications only from the FPR. In addition to the combined total costs, accumulated social- and healthcare costs and medication costs were examined separately. Costs were estimated from the payer’s perspective, including all direct health and social care costs.

Baseline characteristics

Patients’ age and gender were obtained from the EHRs. Concordant and discordant diseases that occurred prior to the T2D diagnosis date were defined from the EHRs (S2 Table). Micro- and macrovascular comorbidities occurring prior to T2D diagnosis were identified utilizing the same definitions as for the outcome (S1 Table). Statin (ATC code C10A) use was identified from the FPR.

For body mass index (BMI), blood pressure, fasting plasma glucose (FPG), and creatinine, measurements occurring most closely to the T2D diagnosis date measured before or within 3 months from the T2D diagnosis, and for low-density lipoprotein (LDL), measurements occurring most closely to the T2D diagnosis date measured before or within 1 month from the diagnosis were extracted from the EHRs. Based on these definitions, indicators of having LDL, BMI, FPG, blood pressure, creatinine and HbA1c measured at the baseline were created. In addition, estimated glomerular filtration rate (eGFR) was calculated with the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation [11] based on creatinine, age, and gender. eGFR was further classified into normal/high (stage I, >90 ml/min/1.73 m2), mildly decreased (stage II, 60–90 ml/min/1.73 m2), mildly to severely decreased (stage III, 30–59 ml/min/1.73 m2) and severely decreased/kidney failure (stage IV, <30 ml/min/1.73 m2).

Diabetes medication use during the follow-up

Diabetes medication use was examined from the FPR for 7 years since T2D diagnosis. The following subgroups of medication use were formed: users of metformin (ATC code A10BA02) only, users of metformin and other oral antidiabetic medications (A10BA02 + other A10B), users of only other diabetes medications than insulin or metformin (A10B excluding A10BA02 and A10A), users of insulin and oral antidiabetic medication (including metformin) (A10A + A10B), and users of insulin (A10A) only. Patients who died were taken into account in the calculation of prevalence of diabetes medication use.

Statistical analyses

Patients were clustered into distinct trajectories of glycaemic control with latent class growth analysis (LCGA) [1215]. Models were fitted iteratively with 1–5 classes and varying shapes for the trajectories (linear, quadratic, cubic). LCGAs were estimated with full-information maximum likelihood and missing data were not imputed but all available data were used. The method assumes no variation within the trajectories. Selection of the best model was done utilizing information from fit indices and classification performance of the model as well as clinical interpretation of the trajectories. Bayesian Information Criteria (BIC) was utilized to compare models with lower values indicating better fit. Low-Mendel-Rubin likelihood ratio test with P value <0.05 was considered to indicate better fit for a n-class model than for n–1-class model. Entropy was used to guide in the classification accuracy of the model with higher values indicating better classification. Small classes with <2% of population were not accepted to have large enough groups of at least 30 patients. LCGAs were performed using Mplus [16] Version 8.

Differences in baseline characteristics between the identified trajectories were examined with one-way ANOVA or nonparametric Kruskal-Wallis test for continuous variables and chi squared test for categorical ones. Association of baseline characteristics with trajectory membership was examined with multinomial logistic regression model. Cox regression models were utilized to describe incidence of deaths and micro- and macrovascular complications by the trajectories. Models were adjusted for age, sex, concordant and discordant diseases, and microvascular and macrovascular comorbidities at baseline. Calendar year- and trajectory-specific healthcare costs were weighted with Kaplan-Meier sample average estimator [17] to account for differences between trajectories in censoring due to death in calculation of adjusted cumulative costs during 2014–2017. Further, non-parametric bootstrapping generating 1000 bootstrap samples with replacement was applied to estimate mean differences in adjusted cumulative healthcare costs with 95% confidence intervals (CIs). Analyses were conducted with SAS version 9.4 (SAS Institute Inc., Cary, North-Carolina, USA). Two-tailed P values < 0.05 were considered statistically significant.

Ethics statement

Use of the data was approved by the Ethics Committee of the Northern Savonia Hospital District (diary number 81/2012) that considers all applications within its university hospital catchment area. The study protocol was also approved by the register administrator, Siun sote—Joint municipal authority for North Karelia social and health services. A separate permission to link data on medication purchases and special reimbursements was achieved from the Social Insurance Institute (diary number 110/522/2018). We utilized only de-identified register-based data and individuals in the registers were not contacted. Thus, no written consent from the patients was required.

Results

In total, 1647 T2D patients diagnosed in 2011–2012 were identified. After removing patients without HbA1c measurements during the follow-up, 1540 (93.5%) patients remained with 14,173 HbA1c measurements. Mean (SD) age of the patients was 64.9 (12.7) years and 46.5% were female (Table 1). 935 (60.7%) had HbA1c measured ±3 months of T2D diagnosis with an average HbA1c of 48.5 (16.5) mmol/mol.

Table 1. Baseline characteristics and variables describing treatment monitoring during the follow-up overall and by the estimated HbA1c trajectories.

Overall cohort Stable, adequate glycaemic control Slowly deteriorating glycaemic control Rapidly deteriorating glycaemic control Late diagnosed patients P value
n (%) 1540 (100.0) 1058 (68.7) 340 (22.1) 95 (6.2) 47 (3.1)
Baseline characteristics
 Female, n (%) 716 (46.5) 514 (48.6) 154 (45.3) 32 (33.7) 16 (34.0) 0.010
 Mean age (SD), years 64.9 (12.7) 65.3 (12.2) 65.5 (13.4) 61.1 (14.2) 60.9 (14.2) 0.002
 FPG measured, n (%) 1122 (72.9) 782 (73.9) 246 (72.4) 73 (76.8) 21 (44.7) <0.001
  Mean FPG (SD), mmol/l 7.6 (2.3) 7.0 (1.4) 8.2 (2.6) 10.5 (4.1) 12.6 (4.1) <0.001
 HbA1c measured, n (%) 935 (60.7) 608 (57.5) 234 (68.8) 60 (63.2) 33 (70.2) 0.001
  Mean HbA1c (SD), mmol/mol 48.5 (16.5) 40.8 (4.8) 47.6 (6.8) 56.0 (11.0) 95.2 (15.1) <0.001
 Blood pressure measured, n (%) 526 (34.2) 382 (36.1) 108 (31.8) 29 (30.5) 7 (14.9) 0.012
 LDL measured, n (%) 1016 (66.0) 721 (68.2) 207 (60.1) 69 (72.6) 19 (40.4) <0.001
  Mean LDL (SD), mmol/l 3.0 (1.0) 3.0 (0.9) 3.0 (1.0) 3.3 (1.0) 3.5 (1.3) 0.008
 HbA1c, LDL and blood pressure measured, n (%) 302 (19.6) 210 (19.9) 70 (20.6) 18 (19.0) 4 (8.5) 0.268
 Creatinine measured, n (%) 1286 (83.5) 889 (84.0) 282 (82.9) 79 (83.2) 36 (76.6) 0.588
  Mean creatinine (SD), μmol/l 76.7 (35.3) 76.9 (39.2) 76.5 (21.9) 79.9 (32.8) 66.4 (21.2) 0.289
  Classified eGFR (ml/min/1.73 m2)
   Stage IV (<30) 8 (0.6) 7 (0.8) 1 (0.4) 0 0 NA
   Stage III (30–59) 120 (9.3) 73 (8.2) 31 (11.0) 11 (13.9) 5 (13.9) 0.219
   Stage II (60–90) 518 (40.3) 369 (41.5) 116 (41.1) 26 (32.9) 7 (19.4) 0.020
   Stage I (>90) 640 (49.8) 440 (49.5) 134 (47.5) 42 (53.2) 24 (66.7) 0.446
 BMI measured, n (%) 673 (43.7) 487 (46.0) 144 (42.4) 33 (34.7) 9 (19.1) 0.001
  Mean BMI (SD), kg/m2 31.0 (6.1) 30.8 (5.9) 31.3 (6.6) 33.3 (6.3) 28.8 (5.9) 0.083
 Concordant diseases, n (%) 473 (30.7) 337 (31.9) 108 (31.8) 24 (25.3) 4 (8.5) 0.005
 Discordant diseases, n (%) 202 (13.1) 149 (14.1) 36 (10.6) 12 (12.6) 5 (10.6) 0.384
 Prevalence of vascular comorbidities, n (%) 158 (10.3) 108 (10.2) 41 (12.1) 6 (6.3) 3 (6.4) 0.312
  Microvascular, n (%) 39 (2.5) 30 (2.8) 7 (2.1) 2 (2.1) 0 0.574
  Macrovascular, n (%) 132 (8.6) 87 (8.2) 38 (11.2) 4 (4.2) 3 (6.4) 0.127
 Prevalence of vascular comorbidities and concordant diseases, n (%) 475 (30.8) 339 (32.0) 108 (31.8) 24 (25.3) 4 (8.5) 0.004
 Statins, n (%) 646 (41.9) 472 (44.6) 125 (36.8) 39 (41.1) 10 (21.3) 0.002
Variables describing treatment monitoring during the follow-up a
 No. of HbA1c measurements, mean (SD) 8.1 (5.2) 7.2 (4.6) 10.4 (5.5) 10.0 (6.5) 8.2 (6.2) <0.001
 Mean HbA1c (SD) 46.3 (11.6) 41.0 (5.5) 51.0 (9.7) 65.2 (16.8) 60.0 (19.7) <0.001

BMI, body mass index; eGFR, estimated glomerular filtration rate; FPG, fasting plasma glucose; LDL, low-density lipoprotein.

aMean follow-up 2016 days (5.5 years); NA, not available.

Of the fitted LCGAs (S3 Table), four-class cubic model was selected as the best fitting with clinically interpretable results (Fig 1). The extracted trajectories were: patients with “Stable, adequate” (66.1%), “Slowly deteriorating” (24.3%), and “Rapidly deteriorating” glycaemic control (6.2%) as well as “Late diagnosed” patients (3.4%).

Fig 1. Estimated HbA1c trajectories.

Fig 1

Patients in the”Rapidly deteriorating” trajectory were more often male and younger than patients in the”Stable, adequate” trajectory (Table 2). Having HbA1c measured at the time of the T2D diagnosis increased and having LDL measured decreased the odds of belonging to the “Slowly deteriorating” and “Late diagnosed” trajectories compared with the”Stable, adequate” trajectory. Having BMI measured decreased the odds of belonging to the “Rapidly deteriorating” and “Late diagnosed” trajectories compared with the”Stable, adequate” trajectory. Patients in the”Late diagnosed” trajectory had less concordant diseases and patients in the “Slowly deteriorating” trajectory used less statins than patients in the”Stable, adequate” trajectory. Average number of HbA1c measurements during the study period differed between the trajectories (Table 1).

Table 2. Adjusted, multivariate odds ratios (95% confidence intervals) for the association of baseline characteristics with estimated trajectories (n = 1540).

Stable, adequate glycaemic control trajectory (n = 1058) as a reference.

Slowly deteriorating glycaemic control (n = 340) Rapidly deteriorating glycaemic control (n = 95) Late diagnosed patients (n = 47)
OR (95% CI) OR (95% CI) OR (95% CI)
Female 0.82 (0.64–1.06) 0.55 (0.35–0.87) 0.51 (0.27–0.98)
Age, years
 <50 1.00 1.00 1.00
 50–74 0.74 (0.49–1.12) 0.37 (0.21–0.64) 0.48 (0.22–1.06)
 ≥75 1.08 (0.68–1.73) 0.37 (0.18–0.75) 0.55 (0.20–1.52)
HbA1c measured 1.97 (1.47–2.64) 1.38 (0.85–2.24) 3.42 (1.59–7.36)
LDL measured 0.68 (0.51–0.91) 1.46 (0.84–2.55) 0.36 (0.19–0.72)
Blood pressure measured 0.84 (0.62–1.13) 0.99 (0.59–1.66) 0.52 (0.22–1.25)
Creatinine measured 0.90 (0.61–1.34) 0.90 (0.46–1.77) 1.02 (0.44–2.41)
BMI measured 0.95 (0.71–1.27) 0.54 (0.33–0.90) 0.39 (0.17–0.86)
Statins 0.75 (0.58–0.98) 0.91 (0.58–1.43) 0.51 (0.24–1.07)
Concordant diseases or micro- or macrovascular comorbidities 1.07 (0.81–1.41) 0.84 (0.50–1.39) 0.26 (0.09–0.74)
Discordant diseases 0.75 (0.50–1.12) 1.05 (0.54–2.01) 1.06 (0.39–2.87)

BMI, body mass index; CI, confidence interval; LDL, low-density lipoprotein; OR, odds ratio.

Diabetes medication use differed between the trajectories and the content of medication structure changed during the 5-year follow-up (Fig 2, S4 Table). In the “Stable, adequate” trajectory, half of the patients did not purchase diabetes medications during the follow-up and metformin use was most frequent. Among patients in the “Rapidly deteriorating” trajectory, treatments were clearly intensified throughout the follow-up. Use of insulin was the most frequent in the “Late diagnosed” trajectory with almost half of the patients using it, remaining stable over the follow-up. Non-use of diabetes medications decreased in every trajectory during the follow-up.

Fig 2. Prevalence of diabetes medication use by HbA1c trajectories.

Fig 2

OAD, oral antidiabetic drugs or GLP-1 analogues (incl. metformin, sulfonylureas, combinations of oral blood glucose lowering drugs, glitazones, DPP-4 inhibitors, glinides, GLP-1 analogues, and SGLT2 inhibitors).

In total, 226 deaths and 318 complications (125 microvascular and 237 macrovascular events) occurred during the follow-up (Fig 3, S1 and S5 Tables). Incidence rate of death after T2D diagnosis was 2.2 (95% CI 1.9–2.6) events per 100 person-years among “Stable, adequate” trajectory, increasing to 3.2 (2.4–4.0) in “Slowly deteriorating”, 4.7 (3.1–6.9) in “Rapidly deteriorating” and 5.2 (2.9–8.7) in the “Late diagnosed” trajectories. Hazard ratio for death was 2.86 (95% CI 1.60–5.12) in the “Late diagnosed” and 2.49 (95% CI 1.61–3.87) in the “Rapidly deteriorating” trajectory compared with the “Stable, adequate” trajectory (S6 Table). Similarly, incidence rate of micro- and macrovascular complications after T2D diagnosis was 3.5 (95% CI 3.0–4.0) events per 100 person-years among “Stable, adequate” trajectory, increasing to 5.1 (4.1–6.2) in “Slowly deteriorating”, 5.5 (3.6–8.1) in “Rapidly deteriorating” and 7.3 (4.3–11.8) in the “Late diagnosed” trajectories. Hazard ratio for complications was 3.35 (95% CI 1.96–5.72) in the “Late diagnosed”, 2.28 (95% CI 1.51–3.46) in the “Rapidly deteriorating”, and 1.39 (95% CI 1.08–1.79) in the “Slowly deteriorating” compared with the “Stable, adequate” trajectory.

Fig 3. Cumulative incidences of health outcomes by the four HbA1c trajectories.

Fig 3

(A) Cumulative incidence of death since T2D diagnosis. (B) Cumulative incidence of combined event of micro- and macrovascular complications since T2D diagnosis.

Accumulated total medication costs differed between the trajectories while no difference was seen in accumulated total costs or accumulated social- and healthcare costs (Table 3, S6 Table). However, patients in the “Late diagnosed” trajectory had lower outpatient visit costs in both primary and specialized healthcare than patients in the “Stable, adequate” trajectory. In addition, patients in the “Stable, adequate” trajectory had lower accumulated diabetes medication costs than patients in other trajectories.

Table 3. Absolute accumulated 4-year per-person healthcare costs (euros, €) for the Stable, adequate glycaemic control trajectory.

Mean differences in accumulated 4-year per-person healthcare costs between HbA1c trajectories with bootstrapped 95% confidence intervals (Stable, adequate glycaemic control trajectory as a reference). All costs are Kaplan-Meier sample average adjusted and calculated for 2014–2017.

Accumulated 4-year per-person costs, € Mean differences in accumulated 4-year per-person costs (€) compared with Stable, adequate glycaemic control trajectory (bootstrapped 95% CI)
Trajectory Stable, adequate glycaemic control Slowly deteriorating glycaemic control Rapidly deteriorating glycaemic control Late diagnosed patients
n (%), alive on Jan 1, 2014 1032 (68.7) 335 (22.1) 90 (6.2) 44 (3.1)
Accumulated total costs 15 341 3882 (-899 to 8872) 1021 (-4122 to 6522) 2817 (-7416 to 18 669)
 1. Accumulated social- and healthcare costs 12 124 3192 (-1306 to 8015) -107 (-4749 to 5156) 1379 (-8443 to 17 069)
  i) Primary healthcare costs 3006 599 (-423 to 1847) 478 (-952 to 2188) -534 (-1843 to 1432)
   -Outpatient visits 1659 54 (-187 to 295) 79 (-446 to 696) -637 (-949 to -302)
   -Inpatient visits 1347 546 (-421 to 1739) 398 (-734 to 1903) 103 (-1094 to 1910)
  ii) Specialized healthcare costs 4543 793 (-1110 to 3184) 48 (-2023 to 2522) -1550 (-3328 to 411)
   -Outpatient visits 1920 -98 (-762 to 551) -435 (-1134 to 312) -1035 (-1664 to -418)
   -Inpatient visits 2622 891 (-679 to 3016) 483 (-1246 to 2745) -515 (-1943 to 1094)
  iii) Nursing home care costs 1200 1767 (-106 to 4035) -396 (-1585 to 1055) -750 (-1691 to 456)
  iv) Home care costs 3376 32 (-2500 to 2629) -236 (-3354 to 3446) 4213 (-3541 to 17 773)
 2. Accumulated medication costs 3216 690 (2 to 1486) 1128 (178 to 2211) 1437 (162 to 2805)
  i) Accumulated diabetes medication costs 363 795 (649 to 959) 2081 (1542 to 2657) 2046 (1283 to 2884)

Discussion

Based our study results, four trajectories among incident Finnish T2D patients were identified with data-driven technique to classify patients following similar HbA1c patterns over time. Identified trajectories had clinical relevance in terms of diabetes medication use, incidence of diabetes complications and deaths as well as accumulation of medication costs during the study. Monitoring of HbA1c and LDL at the time of T2D diagnosis differed between the trajectories. To our knowledge, novelty of the present study is in the examination of direct long-term costs of social- and healthcare by the trajectories among incident T2D patients in Finland. Due to the restricted healthcare resources, it is crucial to understand the costs associated with the trajectories to give reasons for the most effective care of T2D to avoid costly complications.

Trajectories estimated in this study resemble those identified in previous publications. Typically, 2–5 trajectories have been identified among incident T2D populations [1821]. More studies have been conducted among prevalent T2D patients [2227]. Generally, comparison with these previous studies is complicated because of differences in statistical methods used to identify trajectories, follow-up times as well as calendar times: new medications have entered the markets between the studies intensifying the treatment of T2D. In addition, as care and treatment of diabetes varies between geographical areas, and even within one country depending on healthcare providers in addition to differences between populations, results may not be universally generalizable. To note, also clustering based on patient characteristics (age, HbA1c, BMI, homeostasis model assessment (HOMA2) estimates of β-cell function and insulin resistance) at the time of diagnosis in general diabetic population has been examined [28] instead of longitudinal clustering based on HbA1c solely.

Patients in the “Late diagnosed” trajectory had less concordant diseases and measurements of LDL than patients in the “Stable, adequate” trajectory at the time of T2D diagnosis. In addition, having HbA1c measured closely to the T2D diagnosis date increased the odds to be late diagnosed T2D patient compared with the “Stable, adequate” trajectory. These indicate that T2D diagnosis may be delayed due to nonexistence of other diseases or healthcare problems requiring regular controls. In the previous studies [1821], younger age, male sex, and higher BMI have been found to be associated with poorer glycaemic control trajectories similarly to our study.

To examine clinical relevance of the identified trajectories, we described incidences of all-cause deaths and diabetes complications as well as accumulation of healthcare costs by the trajectories. Our results are in line with a systematic review [29] of studies utilizing longitudinal HbA1c clustering and examining their associations with outcomes reporting poor glycaemic control to be associated with the risk of micro- and macrovascular events and mortality. HbA1c may not reflect only treatment response but also the underlying severity of the disease [28]. Late diagnosed patients used most often insulin reflecting more severe disease, seen also as higher incidence of complications during the 7-year follow-up despite of reaching good glycaemic control 2–3 years after the diagnosis. We observed that accumulated medication costs differed between the trajectories and poorer glycaemic control was associated with increased costs which could be considered as a natural consequence of more severe disease, treatment intensification and incidence of complications.

Medical treatments have a central role in T2D care. Newer diabetes medications, i.e., SGLT2 inhibitors and GLP-1 receptor agonists, were reported to lower the risk of cardiovascular events and subsequent mortality during the 2010’s [3, 4]. Use of these medications, especially SGLT2 inhibitors, started to rise in 2016 in Finland. In our cohort, these medications were most often used by patients from the “Rapidly deteriorating” trajectory (S4 Table). In addition, increased use was observed within the “Late diagnosed” and “Slowly deteriorating” trajectories. When examining the risk of micro- and macrovascular complications and mortality, it seems that our analyses may, therefore, underestimate the underlying risk caused by poorer glycaemic control when compared with the “Stable, adequate” trajectory. Of the complications studied in this study, chronic kidney disease may have been somewhat underdiagnosed leading to underestimation of microvascular diseases at baseline and during the follow-up. As the disease is highly related to the mortality risk, residual confounding in mortality analyses is possible although the analyses accounted for end-stage renal disease.

Strengths of our study are inclusion of all patients with T2D diagnosed in 2011–2012 in the North Karelia region. In addition, all the municipalities in North Karelia use the same regional laboratory and the same standardized methods for HbA1c testing. Due to the Finnish public healthcare system available for all inhabitants, we were able to continuously follow patients’ public healthcare service use until death, moving outside of the study area, or administrative study end of the study. Utilization of register-based data allowed us to avoid recall bias, too. Validation studies of the national Finnish Care Register for Health Care, which also includes EHRs from the North Karelia region, have reported positive predictive values of 87–97% for hospital discharges of vascular diseases [30]. Compared with the previous studies, we were also able to examine economic outcomes in addition to health outcomes.

Our study has some limitations. We had a relatively small sample of incident T2D patients stratified to trajectories which limits the conclusions we can draw from the results due to wide confidence intervals especially in the smallest trajectories, although point estimates show some interesting tendencies. For this, we were also unable to conduct an internal validation study. Although we ascertained the timing of T2D diagnosis observed in the EHRs with data from the SII, the date of diagnosis may not be accurate. We had scarce data on laboratory measurements at the baseline meaning that we could not examine the effect of some important factors, such as LDL or BMI, as such on trajectory probability. In addition, according to our previous study, younger patients with T2D were poorly monitored compared with older patients in the North Karelia region [31]. We did not incorporate information on the diabetic medications in use to our analyses but focused solely on HbA1c levels and associated outcomes. However, differences are reported between medications in their efficacy to manage HbA1c [3, 4].

Interpretation of costs should be cautious. Cumulative healthcare costs did not include costs from the first 1–2 years of follow-up since diagnosis. However, as depicted from Fig 3B, differences in the incidence of micro- and macrovascular complications between the trajectories were already apparent at that time possibly causing some bias, especially to specialized healthcare cost analyses. Medication costs did not include costs of medications used during a long-term (>90 days) hospital stays because they are not registered in the FPR. Furthermore, we did not have information on patients using private healthcare services. However, as persons utilizing these services are not likely to use only private healthcare services in the care of T2D due to, for example, economic issues, this is not a big concern. Finally, we aimed to describe the incidence of complications and deaths by the trajectories. Thus, the results presented do not imply causal relationships since trajectories and incidence of the events were measured simultaneously.

Data-driven HbA1c trajectories among incident T2D patients may be utilized as a first step towards individualized care. Repeated measurements of HbA1c are available from EHRs enabling monitoring of patients’ HbA1c levels during a longer period than one measurement point that is often used as a basis for treatment decisions. With the results of this study, clinicians are provided with more information to detect and treat patients whose diabetes care may need more attention. In addition, as patients have a central role in managing their diabetes care, demonstrating their future scenarios including risk of complications, death and healthcare costs may motivate them to engage more intensively to diabetes care. These aspects enable more personalized treatment of T2D patients. Further, this study shows the importance of proper diabetes care and glycaemic management for the society via the association between glycaemic control and medication costs. According to the results, special effort should be allocated to early detection of glycaemic control of late diagnosed T2D patients with high HbA1c values at the time of diagnosis and those with rapid deterioration in treatment balance to decrease the incidence of complications and medication costs. In future, this study should be replicated with a larger sample of T2D patients and longer follow-up to examine causal relationships.

Supporting information

S1 Table. Definitions of micro- and macrovascular complications and numbers of outcome events by type of analysis (first event occurrence after type 2 diabetes diagnosis).

(PDF)

S2 Table. Definitions of concordant and discordant diseases at baseline.

(PDF)

S3 Table. Unadjusted latent class growth analyses for HbA1c values measured annually since T2D diagnosis among patients diagnosed in 2011 or 2012 (n = 1540).

(PDF)

S4 Table. Antidiabetic medication use (%) by the trajectories after the T2D diagnosis.

(PDF)

S5 Table. Associations between the estimated HbA1c trajectories and incidence of deaths and diabetes complications.

(PDF)

S6 Table. Average accumulated 4-year per-person healthcare costs (euros, €) by HbA1c trajectories in 2014–2017.

All the costs are adjusted for censoring due to death utilizing Kaplan-Meier sample average estimator.

(PDF)

Data Availability

Access to data is regulated by the European Union and Finnish laws and therefore, sharing of sensitive data is not possible and data are not publicly available. An anonymised version of the data is available for researchers who meet the criteria as required by the European Union and Finnish laws for access to confidential data with a data permit of an appropriate authority. Contact information: aineistoneuvonta@siunsote.fi for requests from the Siun sote—Joint municipal authority for North Karelia social and health services and tietoaineistot@kela.fi for data requests from the Social Insurance Institute.

Funding Statement

This study was partly supported by the Social Insurance Institution of Finland (https://www.kela.fi/web/en, diary number 110/522/2018), Finnish Diabetes Association (https://www.diabetes.fi/en/finnish_diabetes_association), the Research Committee of the Kuopio University Hospital Catchment Area for the State Research Funding (https://www.psshp.fi/web/en, project QCARE, Joensuu, Finland), and the Strategic Research Council at the Academy of Finland (https://www.aka.fi/en/strategic-research/, project IMPRO, 312703). The funders had no role in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript. The specific roles of these authors are articulated in the ‘author contributions’ section.

References

  • 1.Type 2 diabetes. Current Care Guidelines. Working group set up by the Finnish Medical Society Duodecim, the Finnish Society of Internal Medicine and the Medical Advisory Board of the Finnish Diabetes Society. Helsinki: The Finnish Medical Society Duodecim, 2020 (referred August 6, 2020). https://www.kaypahoito.fi/hoi50056#K1 (22 April 2022, date last accessed).
  • 2.Arnett DK, Blumenthal RS, Albert MA, Buroker AB, Goldberger ZD, Hahn EJ, et al. 2019 ACC/AHA Guideline on the Primary Prevention of Cardiovascular Disease: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines [published correction appears in Circulation. 2019;140(11): e649-e650] [published correction appears in Circulation. 2020;141(4): e60]. Circulation. 2019;140: e596–e646. doi: 10.1161/CIR.0000000000000678 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Cosentino F, Grant PJ, Aboyans V, Bailey CJ, Ceriello A, Delgado V, et al. 2019 ESC Guidelines on diabetes, pre-diabetes, and cardiovascular diseases developed in collaboration with the EASD. Eur Heart J. 2020;41: 255–323. doi: 10.1093/eurheartj/ehz486 [DOI] [PubMed] [Google Scholar]
  • 4.Buse JB, Wexler DJ, Tsapas A, Rossing P, Mingrone G, Mathieu C, et al. 2019 update to: Management of hyperglycaemia in type 2 diabetes, 2018. A consensus report by the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD). Diabetologia. 2020;63: 221–228. doi: 10.1007/s00125-019-05039-w [DOI] [PubMed] [Google Scholar]
  • 5.Coyle D, Buxton MJ, O’Brien BJ. Stratified cost-effectiveness analysis: a framework for establishing efficient limited use criteria. Health Econ. 2003;12: 421–427. doi: 10.1002/hec.788 [DOI] [PubMed] [Google Scholar]
  • 6.Wikström K, Toivakka M, Rautianen P, Tirkkonen H, Kivinen P, Laatikainen T. The effect of the integration of health services on health care usage among patients with type 2 diabetes in North Karelia, Finland. BMC Health Serv Res. 2021:21; 65. doi: 10.1186/s12913-021-06059-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Tirkkonen H, Sikiö M, Kekäläinen P, Laatikainen T. Results of treatment of type 2 diabetes in primary health care (in Finnish with English abstract). Finnish Med J. 2014;34: 2027–2032. [Google Scholar]
  • 8.WHO. International Statistical Classification of Diseases and Related Health Problems 10th revision. http://apps.who.int/classifications/icd10/browse/2010/en. Accessed October 10, 2019.
  • 9.WHO Collaborating Centre for Drug Statistics Methodology, ATC classification index with DDDs, 2019. Oslo, Norway 2018. http://www.whocc.no/atc_ddd_index/. Accessed November 19, 2019.
  • 10.Kapiainen S, Väisänen A, Haula T. Terveyden- ja sosiaalihuollon yksikkökustannukset Suomessa vuonna 2011 (in Finnish). Finnish Institute for Health and Welfare, report 3/2014. http://www.urn.fi/URN:ISBN:978-952-302-079-5. (20 May 2020, date last accessed).
  • 11.Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF 3rd, Feldman HI, et al. CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration). A new equation to estimate glomerular filtration rate. Ann Intern Med 2009;150: 604–612. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Laird NM, Ware JH. Random-effects models for longitudinal data. Biometrics. 1982;38: 963–974. [PubMed] [Google Scholar]
  • 13.Muthén B, Shedden K. Finite mixture modelling with mixture outcomes using the EM algorithm. Biometrics. 1999;55: 463–469. doi: 10.1111/j.0006-341x.1999.00463.x [DOI] [PubMed] [Google Scholar]
  • 14.Jung T, Wickrama KAS. An introduction to latent class growth analysis and growth mixture modeling. Soc Personal Psychol Compass. 2008;2/1: 302–317. [Google Scholar]
  • 15.Nagin DS, Odgers CL. Group-based trajectory modelling in clinical research. Annu Rev Clin Psychol. 2010;6: 109–138. doi: 10.1146/annurev.clinpsy.121208.131413 [DOI] [PubMed] [Google Scholar]
  • 16.Muthén LK, Muthén BO. Mplus User’s Guide, Seventh Edition ed. Muthén & Muthén: Los Angeles, CA; 1998–2015. [Google Scholar]
  • 17.Gray AM, Clarke PM, Wolstenholme JL, Wordsworth S. Applied Methods of Cost-effectiveness Analysis in Healthcare. Oxford University Press, New York, United States, 2011. [Google Scholar]
  • 18.Rathmann W, Schwandt A, Hermann JM, Kuss O, Roden M, Laubner K, et al. Distinct trajectories of HbA1c in newly diagnosed Type 2 diabetes from the DPV registry using a longitudinal group-based modelling approach. Diabet Med. 2019;36: 1468–1477. doi: 10.1111/dme.14103 [DOI] [PubMed] [Google Scholar]
  • 19.Hertroijs DFL, Elissen AMJ, Brouwers MCGJ, Schaper NC, Köhler S, Popa MC, et al. A risk score including body mass index, glycated haemoglobin and triglycerides predicts future glycaemic control in people with type 2 diabetes. Diabetes Obes Metab. 2018;20(3): 681–688. doi: 10.1111/dom.13148 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Laiteerapong N, Karter AJ, Moffet HH, Cooper JM, Gibbons RD, Liu JY, et al. Ten-year hemoglobin A1c trajectories and outcomes in type 2 diabetes mellitus: The Diabetes & Aging Study. J Diabetes Complications. 2017;31: 94–100. doi: 10.1016/j.jdiacomp.2016.07.023 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Hemo B, Geva D, Shahar DR, Golan R, Heymann AD. Distinct trajectories in HbA1c are associated with different all-cause mortality and morbidity in newly diagnosed patients with type 2 diabetes. Prim Care Diabetes. 2019;10: S1751-9918(19)30350-X. doi: 10.1016/j.pcd.2019.11.009 [DOI] [PubMed] [Google Scholar]
  • 22.Luo M, Tan CS, Lim WY, Chia KS, Tang WE, Tai ES, et al. Association of diabetes treatment with long-term glycemic patterns in patients with type 2 diabetes mellitus: A prospective cohort study. Diabetes Metab Res Rev. 2019;35: e3122. doi: 10.1002/dmrr.3122 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Luo M, Lim WY, Tan CS, Ning Y, Chia KS, van Dam RM, et al. Longitudinal trends in HbA1c and associations with comorbidity and all-cause mortality in Asian patients with type 2 diabetes: A cohort study. Diabetes Res Clin Pract. 2017;133: 69–77. doi: 10.1016/j.diabres.2017.08.013 [DOI] [PubMed] [Google Scholar]
  • 24.Walraven I, Mast MR, Hoekstra T, Jansen AP, van der Heijden AA, Rauh SP, et al. Distinct HbA1c trajectories in a type 2 diabetes cohort. Acta Diabetol. 2015;52: 267–275. doi: 10.1007/s00592-014-0633-8 [DOI] [PubMed] [Google Scholar]
  • 25.Ravona-Springer R, Heymann A, Schmeidler J, Moshier E, Godbold J, Sano M, et al. Trajectories in glycemic control over time are associated with cognitive performance in elderly subjects with type 2 diabetes. PLoS One. 2014;9: e97384 doi: 10.1371/journal.pone.0097384 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Chang HY, Wahlqvist ML, Liu WL, Lee MS, Shin SJ, Li YS, et al. Management trajectories in the type 2 diabetes Integrated Delivery System project in Taiwan: accounting for behavioural therapy, nutrition education and therapeutics. Asia Pac J Clin Nutr. 2014;23: 592–606. doi: 10.6133/apjcn.2014.23.4.06 [DOI] [PubMed] [Google Scholar]
  • 27.Karpati T, Leventer-Roberts M, Feldman B, Cohen-Stavi C, Raz I, Balicer R. Patient clusters based on HbA1c trajectories: A step toward individualized medicine in type 2 diabetes. PLoS One. 2018;13: e0207096. doi: 10.1371/journal.pone.0207096 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Ahlqvist E, Storm P, Käräjämäki A, Martinell M, Dorkhan M, Carlsson A, et al. Novel subgroups of adult-onset diabetes and their association with outcomes: a data-driven cluster analysis of six variables. Lancet Diabetes Endocrinol. 2018;6(5): 361–369. doi: 10.1016/S2213-8587(18)30051-2 [DOI] [PubMed] [Google Scholar]
  • 29.Luo M, Tan KHX, Tan CS, Lim WY, Tai ES, Venkataraman K. Longitudinal trends in HbA1c patterns and association with outcomes: A systematic review. Diabetes Metab Res Rev. 2018;34: e3015. doi: 10.1002/dmrr.3015 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Sund R. Quality of the Finnish Hospital Discharge Register: A systematic review. Scand J Public Health. 2012;40: 505–515. doi: 10.1177/1403494812456637 [DOI] [PubMed] [Google Scholar]
  • 31.Nazu NA, Lindström J, Rautiainen P, Tirkkonen H, Wikström K, Repo T, et al. Maintenance of good glycaemic control is challenging—A cohort study of type 2 diabetes patient in North Karelia, Finland. Int J Clin Pract. 2019;73: e13313. doi: 10.1111/ijcp.13313 [DOI] [PubMed] [Google Scholar]

Decision Letter 0

Dylan A Mordaunt

1 Apr 2022

PONE-D-22-05229Data-driven long-term glycemic control clusters and their associated health and economic outcomes in patients with incident type 2 diabetesPLOS ONE

Dear Dr. Lavikainen,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by May 16 2022 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

  • A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

  • A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

  • An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Dylan A Mordaunt

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at 

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and 

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf.

2. Thank you for stating the following in the Competing Interests section: 

[I have read the journal's policy and the authors of this manuscript have the following competing interests: JM is a founding partner of ESiOR Oy and a board member of Siltana Oy. These companies were not involved in carrying out this research. PL, EA, ML, KJ, HT, PR and TL declare no conflicts of interest.]

Please confirm that this does not alter your adherence to all PLOS ONE policies on sharing data and materials, by including the following statement: "This does not alter our adherence to  PLOS ONE policies on sharing data and materials.” (as detailed online in our guide for authors http://journals.plos.org/plosone/s/competing-interests).  If there are restrictions on sharing of data and/or materials, please state these. Please note that we cannot proceed with consideration of your article until this information has been declared. 

Please include your updated Competing Interests statement in your cover letter; we will change the online submission form on your behalf.

3. We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions. 

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially identifying or sensitive patient information) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. Please see http://www.bmj.com/content/340/bmj.c181.long for guidelines on how to de-identify and prepare clinical data for publication. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

4. We note that you have included the phrase “data not shown” in your manuscript. Unfortunately, this does not meet our data sharing requirements. PLOS does not permit references to inaccessible data. We require that authors provide all relevant data within the paper, Supporting Information files, or in an acceptable, public repository. Please add a citation to support this phrase or upload the data that corresponds with these findings to a stable repository (such as Figshare or Dryad) and provide and URLs, DOIs, or accession numbers that may be used to access these data. Or, if the data are not a core part of the research being presented in your study, we ask that you remove the phrase that refers to these data.

5.  Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

Additional Editor Comments:

Thank you for your submission. One of our reviewers has recommended rejection, however the rational is due to a perceived lack of novelty. PLoS One criteria for publication does not include novelty, but rather originality of work. The work does appear to be original. With regards to novelty, the reviewer makes suggestions for improvement that I think the authors should consider. It's also worth considering whether this is novel in the context of the Finish health system and society, since economic outcomes in particular would be health system and society specific. This is the reason I have not receommended rejecting the manuscript. My major area of comment is that any perceived novelty can and should be made clearer, though I've opted to call this a minor rather than major revision.

With specific regards to the criteria for publication:

1. The study appears to present the results of original research.

2. Results do not appear to have been reported elsewhere.

3. Experiments, statistics, and other analyses are performed to a high technical standard and are described in sufficient detail. One reviewer has suggested the authors consider further stratification. I also suggest considering making it clear (e.g. in the title) that this is within the context of Finland, though there may well be generalisable elements or insights.

4. Conclusions are presented in an appropriate fashion and are supported by the data.

5. The article is presented in an intelligible fashion and is written in standard English. Some minor suggestions are made by one of the reviewers, though to a technical English reader, this could affect the perceived credibility of the work since the English is otherwise of a good standard.

6. The research meets all applicable standards for the ethics of experimentation and research integrity.

7. The article could be refined against an appropriate observational study checklist from the Equator Network, such as one of the STROBE checklists. This adds rigour to the reporting. The checklist should be attached to the resubmission as it also adds value to the submission in terms of later meta-research on the resulting publication.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: I Don't Know

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: a useful work which confirms already accepted knowledge regarding the nature and natural history of diabetes, would have been add on to currently known data if there was any stratifications of the drug used and if there was any impact of the class of drugs especially new classes with presumed additive benefits on cardiovascular outcomes rather than just focusing on the progression of the disease

Reviewer #2: Piia Lavikainen et al have applied a clustering strategy to study patients with newly diagnosed type 2 diabetes and examined the association of these groups with heath outcomes including all-cause death and incident micro- and macrovascular complications.

4 groups were identified using this latent-class growth analysis clustering approach : “Stable, adequate” (large majority, 2/3 of the study population), “Slowly deteriorating” (second largest, one fourth of the population) and 2 smal lgroups “Rapidly deteriorating” glycaemic control (6.2%) and “Late diagnosed” patients (3.4%). Most of the interest is focused on these 2 last groups.

The paper is interesting but an effort should be made to facilitate direct access to salient clinical results

General comment

I see A1c not only as a biomarker of treatment efficacy but also as a possible biomarker indicative of severity of the disease as used in Ahlquist intial clustering work (Lancet DE). In addition the response of A1c cannot be seen as a universal response to drugs since diabetes drugs are limited to 8 large classes in patients with T2D (SU/glinides ; TZDs; Metformin; DPP4-Is ; SGLT2-Is; AG-Is; GLP1RA; insulin) although AG-Is do not appear in the current paper.

Table 1

The authors should present data on eGFR and not only serum creatinine values.

Do the authors have access to Fasting plasma glucose at diagnosis ? it would help to confront FPG and A1c.

Table 3 needs to be more directly self-comprehensive as the amount should correspond to yearly expenses (please clarify this)

I am wondering how to adjust on baseline characteristics which are different between the 4 selected groups, particularly age and sex, regarding the effect of each group on complications. This would be appreciated.

The discussion is interesting and could however be shortened.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

PLoS One. 2022 Jun 1;17(6):e0269245. doi: 10.1371/journal.pone.0269245.r002

Author response to Decision Letter 0


25 Apr 2022

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

________________________________________

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: I Don't Know

________________________________________

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

________________________________________

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

________________________________________

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: a useful work which confirms already accepted knowledge regarding the nature and natural history of diabetes, would have been add on to currently known data if there was any stratifications of the drug used and if there was any impact of the class of drugs especially new classes with presumed additive benefits on cardiovascular outcomes rather than just focusing on the progression of the disease

Our response: We thank the Reviewer for the comments. However, we would prefer not to conduct additional stratified analyses. First, diabetes medication use changes over time as is seen from figure 2 and table S4, which also show how medication use modifies HbA1c trajectories. Modeling such a time-varying factor that is also associated with HbA1c would be challenging in this study and we feel that stratified analyses would require different study design. Second, we have submitted another study where we have examined changes in diabetes medication use and its association with treatment outcomes.

Reviewer #2: Piia Lavikainen et al have applied a clustering strategy to study patients with newly diagnosed type 2 diabetes and examined the association of these groups with heath outcomes including all-cause death and incident micro- and macrovascular complications.

4 groups were identified using this latent-class growth analysis clustering approach : “Stable, adequate” (large majority, 2/3 of the study population), “Slowly deteriorating” (second largest, one fourth of the population) and 2 smal lgroups “Rapidly deteriorating” glycaemic control (6.2%) and “Late diagnosed” patients (3.4%). Most of the interest is focused on these 2 last groups.

The paper is interesting but an effort should be made to facilitate direct access to salient clinical results

General comment

I see A1c not only as a biomarker of treatment efficacy but also as a possible biomarker indicative of severity of the disease as used in Ahlquist intial clustering work (Lancet DE). In addition the response of A1c cannot be seen as a universal response to drugs since diabetes drugs are limited to 8 large classes in patients with T2D (SU/glinides ; TZDs; Metformin; DPP4-Is ; SGLT2-Is; AG-Is; GLP1RA; insulin) although AG-Is do not appear in the current paper.

Our response: We have added to the Discussion section on baseline clustering conducted by Ahlqvist et al. as follows: “To note, also clustering based on patient characteristics (age, HbA1c, BMI, homeostasis model assessment (HOMA2) estimates of β-cell function and insulin resistance) at the time of diagnosis in general diabetic population has been conducted[28] instead of longitudinal clustering based on HbA1c solely.”

We have also discussed about the HbA1c being possible indicator of disease severity as follows: “HbA1c may not reflect only treatment response but also the underlying severity of the disease[28]. Late diagnosed patients used most often insulin reflecting more severe disease, seen also as higher incidence of complications during the 7-year follow-up despite of reaching good glycaemic control 2–3 years after the diagnosis.”

AG-Is are not included in the paper as they are not used in Finland. AG-Is are not mentioned in the treatment guidelines of type 2 diabetes in Finland, and there seems to be no AG-I products on the market at the moment.

Table 1

The authors should present data on eGFR and not only serum creatinine values.

Do the authors have access to Fasting plasma glucose at diagnosis ? it would help to confront FPG and A1c.

Our response: We thank the Reviewer for the suggestion and have added eGFR and FPG to the table 1. FPG confronts the observed HbA1c values at baseline. Serum creatinine values were converted to eGFR utilizing the CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration) equation (Levey et al. Ann Intern Med 2009;150:604-612) that accounts for age and gender.

Table 3 needs to be more directly self-comprehensive as the amount should correspond to yearly expenses (please clarify this)

Our response: We thank the Reviewer for commenting on this. We have clarified that the table presents average accumulated 4-year per-person costs over 2014-2017 for the stable cluster and additional costs in other clusters compared with the stable one. We would prefer presenting the accumulated costs along the HbA1c trajectories rather than presenting average yearly costs although they are often reported in cost-effectiveness analyses.

I am wondering how to adjust on baseline characteristics which are different between the 4 selected groups, particularly age and sex, regarding the effect of each group on complications. This would be appreciated.

Our response: We thank the Reviewer for noticing this. Individual level Cox regression models were used to examine adjusted associations of the 4 groups with complications. The models were adjusted for age, sex, concordant and discordant diseases, and microvascular and macrovascular comorbidities of a patient at baseline. Unfortunately, we had not mentioned this in the paragraph of Statistical analyses. Therefore, we have refined the text as follows: “Models were adjusted for age, sex, concordant and discordant diseases, and microvascular and macrovascular comorbidities at baseline.”

The discussion is interesting and could however be shortened.

Our response: We have shortened the discussion by removing, for example, detailed description of HbA1c clusters identified in previous studies.

________________________________________

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

Attachment

Submitted filename: Response to reviewers.docx

Decision Letter 1

Dylan A Mordaunt

18 May 2022

Data-driven long-term glycaemic control trajectories and their associated health and economic outcomes in Finnish patients with incident type 2 diabetes

PONE-D-22-05229R1

Dear Dr. Lavikainen,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Dylan A Mordaunt, MD, MPH, FRACP

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Thank you for your resubmission. This now meets the criteria for publication.

Reviewers' comments:

Acceptance letter

Dylan A Mordaunt

23 May 2022

PONE-D-22-05229R1

Data-driven long-term glycaemic control trajectories and their associated health and economic outcomes in Finnish patients with incident type 2 diabetes

Dear Dr. Lavikainen:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Associate Professor Dylan A Mordaunt

Academic Editor

PLOS ONE

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Supplementary Materials

    S1 Table. Definitions of micro- and macrovascular complications and numbers of outcome events by type of analysis (first event occurrence after type 2 diabetes diagnosis).

    (PDF)

    S2 Table. Definitions of concordant and discordant diseases at baseline.

    (PDF)

    S3 Table. Unadjusted latent class growth analyses for HbA1c values measured annually since T2D diagnosis among patients diagnosed in 2011 or 2012 (n = 1540).

    (PDF)

    S4 Table. Antidiabetic medication use (%) by the trajectories after the T2D diagnosis.

    (PDF)

    S5 Table. Associations between the estimated HbA1c trajectories and incidence of deaths and diabetes complications.

    (PDF)

    S6 Table. Average accumulated 4-year per-person healthcare costs (euros, €) by HbA1c trajectories in 2014–2017.

    All the costs are adjusted for censoring due to death utilizing Kaplan-Meier sample average estimator.

    (PDF)

    Attachment

    Submitted filename: Response to reviewers.docx

    Data Availability Statement

    Access to data is regulated by the European Union and Finnish laws and therefore, sharing of sensitive data is not possible and data are not publicly available. An anonymised version of the data is available for researchers who meet the criteria as required by the European Union and Finnish laws for access to confidential data with a data permit of an appropriate authority. Contact information: aineistoneuvonta@siunsote.fi for requests from the Siun sote—Joint municipal authority for North Karelia social and health services and tietoaineistot@kela.fi for data requests from the Social Insurance Institute.


    Articles from PLoS ONE are provided here courtesy of PLOS

    RESOURCES