ABSTRACT
Aim
To examine trends in potentially avoidable hospitalisations (PAHs) among people with diabetes in Ireland between 2015 and 2024, and how trends differ by sex and diabetes type.
Material and Methods
Using national hospitalisation data from the Republic of Ireland, we estimated rates and trends in PAHs among people with diagnosed type 1 or type 2 diabetes between 2015 and 2024, during which 181 130 PAHs occurred. This included 21 complications, which were used to create a composite measure. We also grouped these into five condition groups including diabetes‐specific, cardiovascular‐related, infection, lower extremity and acute kidney failure. We calculated age‐standardised rates (ASRs) and assessed temporal trends using linear time‐trend models and Joinpoint regression analysis.
Results
ASRs decreased from 105.0 (95% CI: 100.4, 109.2) per 1000 people with diabetes in 2015 to 91.0 (95% CI: 87.7, 94.3) in 2024 with an average decline of 2.5% (95% CI: −3.1, 1.7) per year. ASRs were consistently higher among individuals with type 1 diabetes compared with those with type 2 diabetes (p < 0.001). When examined by condition group, ASRs declined steadily over time for all except those relating to infections, which showed an increase from 2022 onwards.
Conclusion
Although PAHs declined over the study period, rates remained high especially among people with type 1 diabetes. The shifting profile of PAHs towards non‐diabetes‐specific conditions highlights emerging prevention challenges and the need to strengthen primary and community‐based care pathways for people living with diabetes.
Keywords: diabetes complications, observational study, population study, primary care
1. Introduction
Diabetes mellitus is a key contributor to hospital admissions due to a diverse array of acute metabolic and chronic microvascular (e.g., retinopathy, neuropathy, nephropathy) and macrovascular diseases (e.g., cardiovascular and peripheral arterial). Rates of hospitalisations for these conditions have declined in most high income countries over recent decades, but nevertheless remain two to six times as high as persons without diabetes, and many are now increasing again among young and middle‐aged adults [1, 2, 3, 4]. Most complications of diabetes are considered ambulatory care sensitive conditions, whereby escalation to levels that require an emergency hospital admission is thought to be potentially avoidable through timely and appropriate care, including improved medication management within primary care and community settings for those with type 2 diabetes and specialist outpatient care for those with type 1 diabetes [5, 6, 7]. Therefore, they are considered as potentially avoidable hospitalisations (PAHs) or admissions [8] and recognised as being a means of measuring the performance of health systems in delivering primary and preventative care in many high income countries [9].
Most existing studies have focused primarily on PAHs for traditional diabetes‐related complications [10, 11, 12, 13]. However, advances in diabetes management and increased life expectancy are shifting the morbidity profile of people with diabetes. Emerging evidence from several high‐income settings, including the United Kingdom, United States and China, indicates increasing susceptibility to non‐diabetes‐specific conditions, with urinary tract, skin and respiratory infections accounting for a growing proportion of hospitalisations compared with classical diabetes complications [14, 15, 16]. In addition, many studies do not extend into the most recent decade, limiting understanding of how recent developments in diabetes management may have influenced PAH patterns.
Ireland has recently embarked on major health care reform that includes diabetes care and prevention as core components and includes incentivised screening, care management protocols and specialised, short‐term ancillary clinics for people requiring more complex assessment and interventions [17]. These developments highlight the need for a clear understanding of current rates, trends and regional variation in PAHs to guide service planning and resource allocation. However, comprehensive national analyses of the levels, distribution and temporal patterns of PAHs among people with diabetes in Ireland are lacking. Previous Irish research examining avoidable emergency admissions in the general population reported that 4.6% of associated bed‐stay days were attributable to diabetes‐related complications, but this did not provide insights into trends in PAHs among people living with diabetes [18]. Generating robust, diabetes‐specific national data is also relevant internationally, where comparable evidence remains limited. In addition, given the changing epidemiology of complications contributing towards diabetes morbidity, improved understanding of conditions contributing towards PAHs is also needed. This study therefore aimed to examine trends in admissions due to conditions that are potentially amenable to prevention among people with diabetes in Ireland.
2. Materials and Methods
2.1. Study Design and Population
We used national hospitalisation data from the Republic of Ireland to estimate rates and trends in cause‐specific diabetes‐related potential avoidable hospitalisations from 2015 to 2024. Our analyses considered cases of preventable hospitalisations, as described below, as the numerator and the Irish population with diagnosed diabetes as the denominator. As described in more detail in Supporting Information S1, this denominator data was estimated using national prevalence figures derived from Ireland's national Diabetic RetinaScreen programme report and the Irish Childhood Diabetes National Register [19, 20]. The national Hospital Inpatient Enquiry (HIPE) database collects information on all day‐case and in‐patient discharges from public hospitals in Ireland and is managed by the national Healthcare Pricing Office (HPO). Demographic, clinical and administrative data is collected including patient age, sex, mechanism of admission, primary diagnosis upon admission, additional diagnoses and where patients were discharged to or if they died while in hospital. Diagnoses are coded using the International Classification of Diseases version 10, Australian Modification (ICD‐10‐AM). Data was pseudonymised prior to being received by the authors. We assembled a serial cross‐sectional dataset to calculate annual rates of hospital discharges for PAHs in patients who had diabetes (ICD‐10‐AM codes E10‐11‐AM) included as either the primary or a secondary diagnosis. We excluded admissions which included a code related to pregnancy (ICD‐10‐AM code O25) or similar.
2.2. Definition of Potentially Avoidable Hospitalisations
Here we use the term potentially avoidable hospitalisations (PAHs), defined as hospital admissions for conditions where timely and adequate ambulatory care, such as early disease management or preventative measures, could have potentially prevented the need for hospitalisation [8]. We included cases who had a potentially avoidable complication recorded as their principal reason for admission and where they were admitted as an emergency (i.e., reason for admission was not elective). Selection of these complications were informed by a review of the literature [7, 14, 21], and consultation with healthcare professionals working within primary and acute care settings in Ireland. During consultation, healthcare professionals were asked ‘Which diabetes‐related conditions can, if managed effectively in primary or community‐based care, avoid escalation to the point of requiring emergency hospitalisation?’. Selected conditions, corresponding ICD‐10‐AM codes and rationale for inclusion can be seen in Table S1. We selected 21 complications for inclusion, which were used to create a composite measure. We also grouped these into five condition groups: diabetes specific (e.g., diabetic ketoacidosis, hyperglycaemia, hypoglycaemia), cardiovascular related (e.g., stroke, myocardial infarction, transient ischaemic attack), infection (e.g., respiratory and urinary tract infections), lower extremity (e.g., ulceration and peripheral artery disease) and acute kidney failure.
2.3. Statistical Analysis
Demographic characteristics including age, sex and diabetes type of hospital admissions were summarised by year using percentages, with statistical significance of differences between variables assessed using Chi‐square tests. Annual crude rates were calculated for each individual complication, and the composite preventable complication measure, using the estimated national diabetes prevalence figures for each year from 2015 to 2024 (as outlined in Supporting Information S1) as the denominator. Crude rates for the composite measure were calculated by diabetes type, sex and age group (0–24, 25–44, 45–59, 60–74, 75+) and expressed as the number of preventable admissions per 1000 people with diabetes per year. The total population for Ireland in 2024 was used as the standard population for age standardised rates (ASRs), which were calculated for the composite measure, and for each of the five condition groups outlined above [22]. Age was categorised into 5‐year bands for age‐standardisation, and ASRs were calculated by diabetes type and sex, and expressed as the number of preventable admissions per 1000 people with diabetes per year. Statistical significance of differences in rates between years was evaluated using two‐sided Z‐tests. Linear time‐trend models were fitted to estimate the average annual change in ASRs, and joinpoint regression models were subsequently applied to identify any significant changes in trend, stratified by diabetes type. Year was considered as the independent variable, while the ASRs of PAHs were used as dependent variables. This model employs a Monte Carlo permutation method to test for statistically significant changes in trend. The modelling process identified the optimal number of joinpoints (up to a maximum of two), representing points in time where a significant change in the direction or magnitude of the trend occurred. For each trend segment identified, the model estimated an annual percentage change (APC) with corresponding 95% confidence intervals (CIs). Statistical significance was assessed at the 0.05 level. Statistical analyses were performed using R, version 4.3.2 (R Foundation for Statistical Computing). Joinpoint regression software was used for joinpoint analysis (version 5.4.0; Statistical Methodology and Applications Branch and Data Modelling Branch, Surveillance Research Program, National Cancer Institute).
2.4. Sensitivity Analysis
Based on the assumption that those with COVID‐19 may be more susceptible to events requiring hospitalisations and that the COVID‐19 pandemic influenced health seeking behaviours, we also calculated ASRs for those without COVID as an additional diagnosis within the HIPE dataset. Sensitivity analysis, excluding ASRs for 2020, was also run for the time‐trend and joinpoint analysis.
2.5. Ethics Statement
The Research Ethics Committee at the Royal College of Surgeons in Ireland provided ethical approval for using the HIPE data. A pseudonymised data extract was obtained and the authors were not able to identify any specific individuals.
3. Results
3.1. General Characteristics
Between 1 January 2015 and 31 December 2024, there were 1 248 570 hospital admissions for people with a diagnosis of type 1 or type 2 diabetes. Of these, 41% (n = 512 949) were emergency admissions of which 23% (n = 181 130) involved diagnoses indicating it was a PAH. Table 1 summarises the characteristics of admissions and the number of admissions per year. The characteristics of admissions did not differ significantly from those of emergency admissions not considered potentially avoidable (Table S2). Most characteristics were similar over time, with a higher proportion of the cohort being male (57% in 2015 to 60% in 2024, p < 0.001) and having type 2 diabetes (86% in 2015 to 88% in 2024, p < 0.001).
TABLE 1.
Characteristics of potentially avoidable admissions 2015–2024.
| 2015 (N = 15 189) | 2016 (N = 16 804) | 2017 (N = 17 288) | 2018 (N = 17 901) | 2019 (N = 18 516) | 2020 (N = 17 079) | 2021 (N = 17 853) | 2022 (N = 18 393) | 2023 (N = 19 859) | 2024 (N = 22 248) | Overall (N = 181 130) | |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Sex | |||||||||||
| Male | 8767 (57.7%) | 9769 (58.1%) | 10 246 (59.3%) | 10 688 (59.7%) | 11 105 (60.0%) | 10 424 (61.0%) | 10 760 (60.3%) | 11 046 (60.1%) | 11 969 (60.3%) | 13 501 (60.7%) | 108 275 (59.8%) |
| Female | 6422 (42.3%) | 7035 (41.9%) | 7042 (40.7%) | 7213 (40.3%) | 7411 (40.0%) | 6655 (39.0%) | 7093 (39.7%) | 7347 (39.9%) | 7890 (39.7%) | 8747 (39.3%) | 72 855 (40.2%) |
| Diabetes type | |||||||||||
| Type 1 | 2006 (13.2%) | 2242 (13.3%) | 2077 (12.0%) | 2104 (11.8%) | 2097 (11.3%) | 1954 (11.4%) | 2051 (11.5%) | 2179 (11.8%) | 2219 (11.2%) | 2538 (11.4%) | 21 467 (11.9%) |
| Type 2 | 13 183 (86.8%) | 14 562 (86.7%) | 15 211 (88.0%) | 15 797 (88.2%) | 16 419 (88.7%) | 15 125 (88.6%) | 15 802 (88.5%) | 16 214 (88.2%) | 17 640 (88.8%) | 19 710 (88.6%) | 159 663 (88.1%) |
| Age group | |||||||||||
| 0–24 years | 714 (4.7%) | 757 (4.5%) | 638 (3.7%) | 665 (3.7%) | 703 (3.8%) | 647 (3.8%) | 731 (4.1%) | 739 (4.0%) | 735 (3.7%) | 736 (3.3%) | 7065 (3.9%) |
| 25–44 years | 808 (5.3%) | 889 (5.3%) | 907 (5.2%) | 914 (5.1%) | 851 (4.6%) | 821 (4.8%) | 852 (4.8%) | 840 (4.6%) | 891 (4.5%) | 999 (4.5%) | 8772 (4.8%) |
| 45–59 years | 1885 (12.4%) | 2066 (12.3%) | 2129 (12.3%) | 2117 (11.8%) | 2317 (12.5%) | 2226 (13.0%) | 2358 (13.2%) | 2071 (11.3%) | 2241 (11.3%) | 2574 (11.6%) | 21 984 (12.1%) |
| 60–74 years | 5042 (33.2%) | 5573 (33.2%) | 5775 (33.4%) | 6004 (33.5%) | 6130 (33.1%) | 5766 (33.8%) | 5936 (33.2%) | 6001 (32.6%) | 6340 (31.9%) | 7030 (31.6%) | 59 597 (32.9%) |
| 75–84 years | 4654 (30.6%) | 5249 (31.2%) | 5419 (31.3%) | 5518 (30.8%) | 5833 (31.5%) | 5190 (30.4%) | 5374 (30.1%) | 5780 (31.4%) | 6292 (31.7%) | 6975 (31.4%) | 56 284 (31.1%) |
| 85+ years | 2086 (13.7%) | 2270 (13.5%) | 2420 (14.0%) | 2683 (15.0%) | 2682 (14.5%) | 2429 (14.2%) | 2602 (14.6%) | 2962 (16.1%) | 3360 (16.9%) | 3934 (17.7%) | 27 428 (15.1%) |
| Discharged to | |||||||||||
| Self‐discharge | 128 (0.8%) | 141 (0.8%) | 143 (0.8%) | 169 (0.9%) | 155 (0.8%) | 201 (1.2%) | 224 (1.3%) | 244 (1.3%) | 250 (1.3%) | 260 (1.2%) | 1915 (1.1%) |
| Home | 11 178 (73.6%) | 12 498 (74.4%) | 12 668 (73.3%) | 13 073 (73.0%) | 13 731 (74.2%) | 12 804 (75.0%) | 13 136 (73.6%) | 13 432 (73.0%) | 14 260 (71.8%) | 16 096 (72.3%) | 132 876 (73.4%) |
| Nursing home | 1786 (11.8%) | 1894 (11.3%) | 2078 (12.0%) | 2210 (12.3%) | 2121 (11.5%) | 1625 (9.5%) | 1541 (8.6%) | 1897 (10.3%) | 2289 (11.5%) | 2679 (12.0%) | 20 120 (11.1%) |
| Transfer | 1205 (7.9%) | 1347 (8.0%) | 1414 (8.2%) | 1431 (8.0%) | 1535 (8.3%) | 1419 (8.3%) | 1679 (9.4%) | 1641 (8.9%) | 1827 (9.2%) | 1983 (8.9%) | 15 481 (8.5%) |
| Died | 806 (5.3%) | 829 (4.9%) | 896 (5.2%) | 913 (5.1%) | 852 (4.6%) | 945 (5.5%) | 1169 (6.5%) | 1040 (5.7%) | 1042 (5.2%) | 1029 (4.6%) | 9521 (5.3%) |
| Hospice | 28 (0.2%) | 35 (0.2%) | 30 (0.2%) | 39 (0.2%) | 49 (0.3%) | 30 (0.2%) | 39 (0.2%) | 58 (0.3%) | 68 (0.3%) | 81 (0.4%) | 457 (0.3%) |
| Other | 58 (0.4%) | 60 (0.4%) | 59 (0.3%) | 66 (0.4%) | 73 (0.4%) | 55 (0.3%) | 65 (0.4%) | 81 (0.4%) | 123 (0.6%) | 120 (0.5%) | 760 (0.4%) |
3.2. Trends in Rates Over Time
In the total population of people with diabetes, crude rates of PAHs increased from 100.6 [95% CI: 99.1, 102.2] to 102.3 [95% CI: 101.0, 103.6] per 1000 people with diabetes between 2015 and 2024. In contrast, age‐standardised rates (ASRs) decreased from 105.0 [95% CI: 100.4, 109.2] per 1000 people with diabetes in 2015 to 91.0 [95% CI: 87.7, 94.3] in 2024 (Figure 1), indicating divergent temporal trends. Overall, ASRs declined by an average of 2.5% per year (95% CI: −3.1, −1.7), with the reduction occurring primarily between 2015 and 2020 as no significant change was observed thereafter (APC 2.1%; 95% CI: −2.3, 5.3). When examined by diabetes type, crude rates (Tables S3–S5) declined among people with type 1 diabetes (106.9 [95% CI: 102.5, 111.4] to 100.4 [95% CI: 96.7, 104.2]; p < 0.05), but increased among those with type 2 diabetes (99.8 [95% CI: 98.1, 101.4] to 102.5 [95% CI: 101.2, 103.9]; p < 0.05) between 2015 and 2024. Reductions among people with type 1 diabetes were primarily driven by fewer admissions for hypoglycaemia, whereas increases in type 2 diabetes reflected rising admissions for angina, urinary tract infections and respiratory infections. In contrast, ASRs were consistently higher among individuals with type 1 diabetes compared to those with type 2 diabetes (p < 0.001). Although rates of decline did not differ significantly between groups, reductions were only statistically significant among those with type 2 diabetes with rates declining by 2.4% per year [95% CI: 0.4, 4.2; p = 0.02]. Joinpoint regression indicates that for type 1 diabetes, reductions were concentrated between 2015 and 2020 (APC: −5%, 95% CI: −11, −2.6), whereas the reduction among persons with type 2 diabetes decreased consistently across the full study period (APC: –2.3%; 95% CI: −4.5, −0.1). No sex differences in ASRs were observed for the total population with diabetes or for those with type 2 diabetes (Figure 2). However, ASRs among people with type 1 diabetes were significantly higher in women compared with men (Figure 2; Table S6). Following sensitivity analysis, ASRs were marginally higher among those with COVID‐19 with an additional diagnosis compared to those without (Figures S1 and S2). However, the overall pattern of change over time was unchanged. Joinpoint analyses showed similar inflection points, direction of trends and statistical significance to the main analysis, indicating that exclusion of COVID‐19‐related admissions did not materially alter the principal findings or conclusions of the study. Similar findings were observed when the year 2020 was excluded (Figure S3). While omission of 2020 resulted in minor changes in the timing and identification of temporal inflection points, the broader pattern of declining ASRs remained consistent. In particular, rates continued to decline overall and among people with type 2 diabetes, suggesting the main conclusions remained robust even after excluding the first year of the COVID‐19 pandemic.
FIGURE 1.
Age Standardised Rates of potentially avoidable hospitalisations for all people with diabetes, for those with type 1 diabetes only and for those with type 2 diabetes only.
FIGURE 2.
Age standardised rates of potentially avoidable hospitalisations by diabetes type and sex.
3.3. Grouped Conditions
When examined by condition group and for all people with diabetes, ASRs were highest for diabetes specific events, followed by infection, cardiovascular, lower limb and acute kidney failure related events (Figure 3), with declines over time for all except those relating to infections. Joinpoint analysis shows much of these reductions occurred steadily over time, with the exception of infection‐related events which showed an increase from 2022 onwards (Figures S4–S8). However, there were differences in trends when examined by diabetes type and sex. For diabetes specific events, ASRs declined overall by 2.8% per year (95% CI: −1.7%, −3.8%). This decline was observed only in people with type 2 diabetes (−6.6% per year [95% CI: −3.6%, −9.5%]), whereas the change in people with type 1 diabetes was smaller and not statistically significant (−1.2% per year; 95% CI: −2.24%, 0.17%). There was also an overall decline in cardiovascular‐related events, with ASRs declining by 1.6% [95% CI: 0.8%, 2.4%] per year for all people with diabetes. When stratified by diabetes type, the downward trends persisted but neither group reached statistical significance. For events relating to infection, trend analysis showed no significant change over time for the total population or for people with type 1 diabetes, but ASRs did increase by 2.7% per year [95% CI: 1.2, 6.7] in those with type 2 diabetes with much of this increase occurring between 2022 and 2024 with an APC of 21.9% [95% CI: 1.8, 41.5] (Figure S5). For PAHs relating to lower limb conditions, trend analysis indicated a statistically significant annual decline in ASRs of 4.6% [95% CI: −7.5%, −1.6%] per year for all people with diabetes. When examined by diabetes type, ASRs declined in both groups but this reduction was only significant for people with type 2 diabetes (7.5%; 95% CI: −1.5%, −13.1%). Joinpoint analysis (Figure S7) confirmed this downward pattern (APC: 4.5% [95% CI: −4.5, −1.4]) during the study period, with a greater reduction observed in type 2 diabetes compared with type 1 (7.5% vs. 4.9%). No significant changes were observed for hospitalisation due to acute kidney failure (Figure S8). For diabetes‐specific and cardiovascular related events, ASRs were higher in females compared to males but were similar between sexes over time for all other condition groups. However, these findings were accompanied by wide confidence intervals (Tables S7–S11; Figure S9).
FIGURE 3.
Age standardised rates of potentially avoidable hospitalisations by condition group (diabetes specific, cardiovascular related, infection related, lower limb specific and acute kidney failure) by diabetes type.
4. Discussion
Potentially avoidable hospitalisations (PAHs) are widely used as indicators of the effectiveness of primary and community‐based care in managing chronic disease [9]. In this nationally representative analysis of public hospital admissions between 2015 and 2024 in Ireland, we found that while overall ASRs of PAHs declined by 2.5% per year, much of this occurred between 2015 and 2022. In addition, the primary contributors to PAHs have changed, with a reduced contribution from traditional diabetes‐specific complications and a growing contribution from non–diabetes‐specific conditions, particularly respiratory, skin and urinary tract infections. We also found differences across population subgroups as rates remained consistently higher among people with type 1 diabetes compared to those with type 2 diabetes, and higher among women with type 1 diabetes.
Our study provides a novel perspective on trends in PAHs among people with diabetes. Comparison with previous studies is limited by their focus on traditional diabetes‐specific complications, by observation periods that did not extend into the most recent decade and by not accounting for changes in population structures over time [10, 11, 13, 23]. Although declines in diabetes complications (e.g., DKA, Hypoglycaemic events) reported by others are consistent with the reductions we observed in hospitalisation relating to diabetes‐specific events [23], our broader findings contrast with PAH trends from other high‐income countries which are reported to have either plateaued or increased [10, 12]. Trends in PAHs in these countries appear to be driven by increases in lower extremity amputations. These were not included in our definition of PAHs as admissions for these events were typically elective rather than emergency. In addition, amputations typically occur following hospitalisation for an acute event such as peripheral arterial disease, ulceration or infection. Despite the downward trend identified in our analysis, the ASRs we observed remain higher than those reported in other comparable regions, indicating the burden of PAHs among people with diabetes in Ireland is still substantial. For example, in Southern Italy, a study of short‐term diabetes complications, which are most comparable to the diabetes‐specific events examined in our analysis, reported about a 70% reduction in ASRs over 10 years [23]. Despite similar age and sex profile, rates in Italy were substantially lower than our findings, indicating a markedly higher burden of diabetes‐specific PAHs in Ireland. It is also important to highlight the divergence observed in our results between increasing crude rates and declining ASRs among people with type 2 diabetes. This pattern likely reflects population ageing, whereby a growing and older population contributes to rising crude rates despite improvements in age‐specific risk. Therefore, the observed decline in ASRs may mask an increasing absolute burden of PAHs in this population.
A major shift was also observed relating to the changing composition of conditions contributing to PAHs. Although people with diabetes are known to have a higher susceptibility to infections than the general population [16, 24], hospital admissions in this group have traditionally been dominated by events directly related to diabetes management, including hyperglycaemia, hypoglycaemia and cardiovascular complications [25]. We found a decline in admissions due to hyper‐ and hypoglycaemia and a corresponding increase in hospitalisations for conditions not specific to diabetes, including respiratory and urinary tract infections. Similar transitions have been reported in other high‐income countries [14, 15], suggesting a broader epidemiological shift in the burden of complications amenable to prevention efforts among people with diabetes, particularly those with type 2 diabetes. Although this pattern may reflect ageing populations, longer diabetes duration and declining traditional complications, the rising contribution of infections is concerning given the higher risk of adverse outcomes, including mortality, among people with diabetes who develop infections compared with the general population [26]. Further research is needed to clarify the drivers of this shift and to determine the extent to which it reflects changes in population risk, care delivery or both. For cardiovascular, infection‐related, lower limb and acute kidney failure hospitalisations, we did not have access to comparable data from the Irish general population. There is also a limited number of published studies on trends in hospitalisations for these conditions in the Irish general population, making it difficult to assess whether observed trends are diabetes‐specific or reflect broader patterns. However, one Irish study examining emergency admissions among adults aged ≥ 65 years between 2012 and 2016 found that hospitalisations for urinary tract infections were increasing, whereas admissions for angina and congestive heart failure were decreasing [27]. Using national hospitalisation data from 2005 to 2015 another Irish study found that while hospitalisations rates for ischaemic stroke were higher for those with diabetes, there was a decline in both people with and without diabetes [28]. These findings suggest that some of the reductions observed in cardiovascular related PAHs may reflect broader population‐level improvements in cardiovascular prevention, acute management and healthcare delivery, rather than diabetes‐specific changes alone.
Differences in trends were also evident by diabetes type, and to our knowledge, this is the first population‐based study to examine trends in PAHs separately by diabetes type and so the first to show the persistently higher PAH rates among individuals with type 1 diabetes. While this may reflect the often‐greater clinical complexity of this population and their dependence on specialist‐led care, some studies have suggested that performance‐based incentives may inadvertently exacerbate inequalities in chronic disease management, favouring conditions with primary care metrics [29]. In Ireland, structured primary care initiatives such as the Cycle of Care for Diabetes (2015), the Chronic Disease Management Programme (2020) and Enhanced Community Care (2021) have strengthened primary care management of chronic diseases, particularly type 2 diabetes, through incentivised reviews, care planning and structured follow‐up. However, people with type 1 diabetes are largely excluded. This differential policy focus may have contributed to the persistently higher rates of PAHs observed among those with type 1 diabetes, thus underscoring the need to extend structured chronic disease management supports to this population. Nonetheless, these policy initiatives do not fully explain the observed reduction in crude rates among younger cohorts with type 1 diabetes (aged 0–24 years) alongside the increase among middle‐aged adults (aged 45–59 and 60–74 years). These contrasting trends may be partly attributable to previously reported age‐related differences in the uptake of diabetes technologies, which tend to be adopted earlier and more extensively among younger individuals and may offer protection against acute complications [30]. We also extended previous work by assessing sex differences both overall and within diabetes type. In contrast to findings from a study in the United Kingdom [14], we observed no sex differences in overall PAHs or across condition groups. However, disparities were observed among people with type 1 diabetes, where rates for hospitalisations relating to diabetes‐specific, infections and vascular events were nearly twice as high for women compared to men. Although higher rates of diabetic ketoacidosis among women are well documented [31], the elevated rates of infection‐related hospitalisations we identified contrast with findings from other high‐income countries [32] with the exception of urinary tract infections, for which higher rates among women are consistently reported. While a similar pattern for diabetes‐specific and infection related events was observed among males and females with type 2 diabetes, these findings should be interpreted with caution as estimates were accompanied by wide confidence intervals, limiting the precision of observed trends. Although the observed consistency in direction suggests the possibility of sex‐related differences in PAHs regardless of diabetes type, further investigation is needed to confirm definitive disparities especially among those with type 2 diabetes. This is important because higher PAH rates among women may reflect differences in access to timely primary and community‐based care, variations in symptom recognition or care‐seeking behaviour, or inequities in the tailoring of clinical management to women's needs [33]. Such disparities are seldom examined in diabetes research or health‐system performance metrics, yet they have direct implications for the design of prevention strategies and care pathways. This was evident in the current study, where several potentially important explanatory factors were not available for adjustment within the data, including socioeconomic deprivation and regional variation in healthcare access and service provision. These factors may partly explain some of the observed differences by diabetes type and sex as we know they are often associated with differences in diabetes prevalence and development of associated complications [34, 35, 36]. However, development of datasets that have the scope to explore the associations between socioeconomic deprivation and geographical variation in outcomes in people with diabetes in Ireland are needed to facilitate further research in this area.
4.1. Strengths and Limitations
This study has several strengths. First, it is based on national data covering all admissions to public hospitals in Ireland during the study period. Second, to the best of our knowledge this is the first national study to examine differences in PAHs in those with type 1 vs. type 2 diabetes thus providing novel insights. However, this study also has its limitations. First, we do not know if patients received care in the community setting prior to hospitalisation, yet the number of GP encounters influences PAH rates [37]. Second, this study only examines trends in those with diabetes and does not allow for comparison with the general population meaning hospitalisations for many of these conditions may also be increasing in other groups. Third, HIPE data does not include admissions to private hospitals and so these figures may be an underestimation of rates. Fourth, diabetes status was identified from diagnostic coding recorded during hospital admissions. While this may have resulted in some misclassification or under‐ascertainment, diabetes coding in HIPE is linked to hospital reimbursement, which is likely to support relatively high completeness. We also did not examine presence of comorbid conditions which may also increase a person's risk of hospitalisation for selected conditions. Finally, it is important to note that numerator and denominator data were derived from different national sources. Hospital admissions were identified using HIPE discharge data, whereas denominator estimates were based on diagnosed diabetes prevalence derived from Diabetic RetinaScreen programme data and the Irish Childhood Diabetes National Register. These data sources use different approaches to classify diabetes, which may have led to misclassification of diabetes type and contributed to numerator–denominator mismatch, potentially affecting type‐specific rate estimates. In addition, as outlined in more detail in Supporting Information S1 describing the methods used to calculate estimated prevalence of diabetes, a potential limitation of the denominator is that it may underestimate the true prevalence of diabetes in Ireland, which could result in overestimation of PAH rates. However, a recently published study examining diabetes prevalence by age and diabetes type reported an overall diabetes prevalence of approximately 4%, which is consistent with the estimates used in our denominator data [34]. In addition, any underestimation would have remained relatively consistent over time and so it would be expected to influence annual rates in a similar way each year, meaning the observed temporal trends are likely to remain reasonably robust. Furthermore, numerator and denominator estimates were aligned to the same calendar years, with denominator estimates also incorporating age‐ and sex‐specific Central Statistics Office population data, helping to account for changes in the underlying population structure over time.
5. Conclusion
This nationally representative study highlights important trends in PAHs among people with diabetes in Ireland. Overall, ASRs have declined over the study period, but with no significant change observed since 2022. However, increases in specific causes, particularly infection‐related admissions among people with type 2 diabetes, are concerning. Persistent disparities by diabetes type and sex suggest there may be ongoing inequities in access to preventive care. Strengthening community‐based management through health care reforms such as the Enhanced Community Care Programme remains essential, alongside extending structured supports and access to diabetes technologies for individuals with type 1 diabetes.
Funding
This publication has emanated from research conducted with the financial support of Research Ireland under Grant number [22/RP/10091]. The funding body had no role in the design and conduct of the study; collection, management, analysis and interpretation of the data; preparation, review or approval of the manuscript; or decision to submit the manuscript for publication.
Conflicts of Interest
Seamus Sreenan is on the board of directors of Diabetes Ireland and the Blanchardstown Hospital Society and receives no compensation as member of either board of directors. In the last five years, Seamus Sreenan has received honoraria for participation in advisory boards for Abbott Laboratories and Novo Nordisk and for receives honoraria from Eli Lilly for participation in a non‐promotional annual conference. All other authors have no conflicts of interest to declare.
Supporting information
Data S1: Methods used to calculate estimated prevalence of diabetes.
Table S1: Selected potentially avoidable acute complications and ICD‐10 codes.
Table S2: Characteristics of non‐preventable emergency admissions.
Table S3: Crude rates for selected complications per 1000 people with diabetes.
Table S4: Crude rates of potentially avoidable hospitalisations per 1000 people with diabetes.
Table S5: Crude rates of potentially avoidable complications by diabetes type and age group per 1000 people with diabetes.
Table S6: Age standardised rates by diabetes type and sex per 1000 people with diabetes.
Table S7: Age standardised rates by diabetes type and sex per 1000 people with diabetes—Diabetes specific events.
Table S8: Age standardised rates by diabetes type and sex per 1000 people with diabetes—Vascular events.
Table S9: Age standardised rates by diabetes type and sex per 1000 people with diabetes—Infections.
Table S10: Age standardised rates by diabetes type and sex per 1000 people with diabetes—Lower limb.
Table S11: Age standardised rates by diabetes type and sex per 1000 people with diabetes—Acute kidney failure.
Figure S1: Age standardised rates of acute complications per 1000 people with diabetes without COVID as additional diagnosis.
Figure S2: Join point analysis for all complications with and without COVID‐19 as an additional diagnosis.
Figure S3: Join point analysis for all complications excluding year 2020.
Figure S4: Join point analysis for diabetes specific hospitalisations.
Figure S5: Join point analysis for infection related hospitalisations.
Figure S6: Join point analysis for cardiovascular related hospitalisations.
Figure S7: Join point analysis for lower limb related hospitalisations.
Figure S8: Join point analysis for acute kidney failure related hospitalisations.
Figure S9: Age standardised rates of grouped conditions by diabetes and sex.
Pallin J. A., Holman N., McIntyre K., et al., “Trends in Potentially Avoidable Hospitalisations Among People With Diabetes in Ireland; 2015–2024,” Diabetes, Obesity and Metabolism 28, no. 8 (2026): 7161–7170, 10.1111/dom.70925.
Handling Editor: Richard Donnelly
Data Availability Statement
The data used in this study are not publicly available due to restrictions under Irish and European data protection legislation, in order to protect patient privacy. The data were derived from the Hospital In‐Patient Enquiry (HIPE) system, which is managed by the Healthcare Pricing Office (HPO) on behalf of the Health Service Executive (HSE) in Ireland. Access to HIPE data may be granted to researchers with appropriate ethical approvals and who meet the criteria for access to confidential data, subject to approval by the HPO and relevant data governance processes. Requests for access to the data should be directed to the Healthcare Pricing Office. The analytical code used to support the findings of this study may be made available by the corresponding author upon reasonable request, where permitted by data governance restrictions.
References
- 1. Honigberg M. C., Patel R. B., Pandey A., et al., “Trends in Hospitalizations for Heart Failure and Ischemic Heart Disease Among US Adults With Diabetes,” JAMA Cardiology 6, no. 3 (2021): 354–357, 10.1001/JAMACARDIO.2020.5921. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Read S. H., Fischbacher C. M., Colhoun H. M., et al., “Trends in Incidence and Case Fatality of Acute Myocardial Infarction, Angina and Coronary Revascularisation in People With and Without Type 2 Diabetes in Scotland Between 2006 and 2015,” Diabetologia 62, no. 3 (2019): 418–425, 10.1007/S00125-018-4796-7/FIGURES/3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Gregg E. W., Hora I., and Benoit S. R., “Resurgence in Diabetes‐Related Complications,” Journal of the American Medical Association 321, no. 19 (2019): 1867–1868, 10.1001/JAMA.2019.3471. [DOI] [PubMed] [Google Scholar]
- 4. Gregg E. W., Pratt A., Owens A., et al., “The Burden of Diabetes‐Associated Multiple Long‐Term Conditions on Years of Life Spent and Lost,” Nature Medicine 30, no. 10 (2024): 2830–2837, 10.1038/s41591-024-03123-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Feleke B. E., Shaw J. E., and Magliano D. J., “Trends in Rates of Hospitalisation for Infection in People With Diabetes and the General Population,” Diabetic Medicine 41 (2024): e15421, 10.1111/DME.15421. [DOI] [PubMed] [Google Scholar]
- 6. Lin W., Huang I. C., Wang S. L., Yang M. C., and Yaung C. L., “Continuity of Diabetes Care Is Associated With Avoidable Hospitalizations: Evidence From Taiwan's National Health Insurance Scheme,” International Journal for Quality in Health Care 22, no. 1 (2010): 3–8, 10.1093/INTQHC/MZP059. [DOI] [PubMed] [Google Scholar]
- 7. Wharam J. F., Argetsinger S., Lakoma M., Zhang F., and Ross‐Degnan D., “Acute Diabetes Complications After Transition to a Value‐Based Medication Benefit,” JAMA Health Forum 5, no. 2 (2024): E235309, 10.1001/jamahealthforum.2023.5309. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Purdy S., Griffin T., Salisbury C., and Sharp D., “Ambulatory Care Sensitive Conditions: Terminology and Disease Coding Need to Be More Specific to Aid Policy Makers and Clinicians,” Public Health 123, no. 2 (2009): 169–173, 10.1016/J.PUHE.2008.11.001. [DOI] [PubMed] [Google Scholar]
- 9. Kringos D. S., Boerma W. G., Hutchinson A., Van Der Zee J., and Groenewegen P. P., “The Breadth of Primary Care: A Systematic Literature Review of Its Core Dimensions,” BMC Health Services Research 10, no. 1 (2010): 65, 10.1186/1472-6963-10-65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Maung K. K., Mauron C., Gouveia A., and Marques‐Vidal P., “Trends in Potentially Avoidable Hospitalizations for Diabetes in Switzerland, 1998 to 2018: Data From Multiple Cross‐Sectional Studies,” Heliyon 10, no. 23 (2024): e40466, 10.1016/J.HELIYON.2024.E40466. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Manderbacka K., Arffman M., Lumme S., Lehikoinen M., Winell K., and Keskimäki I., “Regional Trends in Avoidable Hospitalisations due to Complications Among Population With Diabetes in Finland in 1996‐2011: A Register‐Based Cohort Study,” BMJ Open 6, no. 8 (2016): e011620, 10.1136/BMJOPEN-2016-011620. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Rubens M., Saxena A., Ramamoorthy V., et al., “Trends in Diabetes‐Related Preventable Hospitalizations in the U.S., 2005–2014,” Diabetes Care 41, no. 5 (2018): e72–e73, 10.2337/DC17-1942. [DOI] [PubMed] [Google Scholar]
- 13. Ramalho A., Lobo M., Duarte L., Souza J., Santos P., and Freitas A., “Landscapes on Prevention Quality Indicators: A Spatial Analysis of Diabetes Preventable Hospitalizations in Portugal (2016–2017),” International Journal of Environmental Research and Public Health 17, no. 22 (2020): 8387, 10.3390/IJERPH17228387. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Pearson‐Stuttard J. C. Y., Bennett J., Vamos E. P., et al., “Trends in Leading Causes of Hospitalisation of Adults With Diabetes in England From 2003 to 2018: An Epidemiological Analysis of Linked Primary Care Records,” Lancet Diabetes and Endocrinology 10 (2022): 46–57, 10.1016/S2213-8587(21)00288-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Wu H., Zhou H., Huang C., et al., “Identifying and Visualising Temporal Trajectories of Hospitalisations for Traditional and Non‐Traditional Complications in People With Type 2 Diabetes: A Population‐Based Study,” Lancet Regional Health – Western Pacific 57 (2025): 101532, 10.1016/j.lanwpc.2025.101532. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Harding J. L., Benoit S. R., Gregg E. W., Pavkov M. E., and Perreault L., “Trends in Rates of Infections Requiring Hospitalization Among Adults With Versus Without Diabetes in the U.S., 2000–2015,” Diabetes Care 43, no. 1 (2020): 106–116, 10.2337/DC19-0653. [DOI] [PubMed] [Google Scholar]
- 17. HSE , “Chronic Disease Management Programme,” (2025), https://www.hse.ie/eng/services/list/2/primarycare/chronic‐disease‐management‐programme/.
- 18. Mcdarby G. and Smyth B., “Identifying Priorities for Primary Care Investment in Ireland Through a Population‐Based Analysis of Avoidable Hospital Admissions for Ambulatory Care Sensitive Conditions (ACSC),” BMJ Open 9, no. 11 (2019): e028744, 10.1136/bmjopen-2018-028744. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Diabetes Ireland , “Irish Childhood Diabetes National Register ‐ Diabetes Ireland: Diabetes Ireland,” (2025), https://www.diabetes.ie/living‐with‐diabetes/child‐diabetes/16867‐2/.
- 20. Health Service Executive , “Diabetic Retina Screening—Diabetic RetinaScreen,” (2026), https://www2.hse.ie/conditions/diabetic‐retina‐screening/.
- 21. Sundmacher L., Fischbach D., Schuettig W., Naumann C., Augustin U., and Faisst C., “Which Hospitalisations Are Ambulatory Care‐Sensitive, to What Degree, and How Could the Rates Be Reduced? Results of a Group Consensus Study in Germany,” Health Policy (New York) 119, no. 11 (2015): 1415–1423, 10.1016/J.HEALTHPOL.2015.08.007. [DOI] [PubMed] [Google Scholar]
- 22. CSO , “Annual Population Estimates,” (2024), https://data.cso.ie/.
- 23. Martino G. D., Di Giovanni P., Cedrone F., et al., “The Burden of Diabetes‐Related Preventable Hospitalization: 11‐Year Trend and Associated Factors in a Region of Southern Italy,” Healthcare (Basel) 9, no. 8 (2021): 997, 10.3390/HEALTHCARE9080997. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Carey I. M., Critchley J. A., Dewilde S., Harris T., Hosking F. J., and Cook D. G., “Risk of Infection in Type 1 and Type 2 Diabetes Compared With the General Population: A Matched Cohort Study,” Diabetes Care 41, no. 3 (2018): 513–521, 10.2337/DC17-2131. [DOI] [PubMed] [Google Scholar]
- 25. AbuHammad G. A. R., Naser A. Y., and Hassouneh L. K. M., “Diabetes Mellitus‐Related Hospital Admissions and Prescriptions of Antidiabetic Agents in England and Wales: An Ecological Study,” BMC Endocrine Disorders 23, no. 1 (2023): 1–16, 10.1186/S12902-023-01352-Z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Carey I. M., Critchley J. A., Chaudhry U. A. R., et al., “Contribution of Infection to Mortality in People With Type 2 Diabetes: A Population‐Based Cohort Study Using Electronic Records,” Lancet Regional Health—Europe 48 (2025): 101147, 10.1016/j.lanepe.2024.101147. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Walsh M. E., Cronin S., Boland F., Ebell M. H., Fahey T., and Wallace E., “Geographical Variation of Emergency Hospital Admissions for Ambulatory Care Sensitive Conditions in Older Adults in Ireland 2012–2016,” BMJ Open 11, no. 5 (2021): e042779, 10.1136/BMJOPEN-2020-042779. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Bambury N., O'Neill K., Buckley C. M., and Kearney P. M., “Trends in Incidence of Ischaemic Stroke in People With and Without Diabetes in Ireland 2005–2015,” Diabetic Medicine 40, no. 11 (2023): e15127, 10.1111/DME.15127;PAGE:STRING:ARTICLE/CHAPTER. [DOI] [PubMed] [Google Scholar]
- 29. Linnane S., Mullarkey S., Kyne E., et al., “Does Pay for Performance Promote Inverse Inequality in Chronic Disease Management?,” Family Practice 42, no. 3 (2025): cmaf025, 10.1093/FAMPRA/CMAF025. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Gajewska K. A., Bennett K., Biesma R., and Sreenan S., “Low Uptake of Continuous Subcutaneous Insulin Infusion Therapy in People With Type 1 Diabetes in Ireland: A Retrospective Cross‐Sectional Study,” BMC Endocrine Disorders 20, no. 1 (2020): 1–8, 10.1186/S12902-020-00573-W/TABLES/2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Farsani S. F., Brodovicz K., Soleymanlou N., Marquard J., Wissinger E., and Maiese B. A., “Incidence and Prevalence of Diabetic Ketoacidosis (DKA) Among Adults With Type 1 Diabetes Mellitus (T1D): A Systematic Literature Review,” BMJ Open 7, no. 7 (2017): e016587, 10.1136/BMJOPEN-2017-016587. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Luk A. O. Y., Wu H., Lau E. S. H., et al., “Temporal Trends in Rates of Infection‐Related Hospitalisations in Hong Kong People With and Without Diabetes, 2001–2016: A Retrospective Study,” Diabetologia 64, no. 1 (2020): 109–118, 10.1007/S00125-020-05286-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Gisinger T., Azizi Z., Alipour P., et al., “Sex and Gender Aspects in Diabetes Mellitus: Focus on Access to Health Care and Cardiovascular Outcomes,” Frontiers in Public Health 11 (2023): 1090541, 10.3389/FPUBH.2023.1090541/PDF. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Valentelyte G., Holman N., James S., et al., “Socioeconomic Disparities in Diabetes Prevalence Among the Population in Ireland,” BMC Public Health 25, no. 1 (2025): 2206, 10.1186/S12889-025-23022-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Leese G. P., Feng Z., Leese R. M., Dibben C., and Emslie‐Smith A., “Impact of Health‐Care Accessibility and Social Deprivation on Diabetes Related Foot Disease,” Diabetic Medicine 30, no. 4 (2013): 484–490, 10.1111/DME.12108. [DOI] [PubMed] [Google Scholar]
- 36. Kurani S. S., Lampman M. A., Funni S. A., et al., “Association Between Area‐Level Socioeconomic Deprivation and Diabetes Care Quality in US Primary Care Practices,” JAMA Network Open 4, no. 12 (2021): e2138438, 10.1001/JAMANETWORKOPEN.2021.38438. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Georgescu V., Green A., Jensen P. B., Möller S., Renard E., and Mercier G., “Primary Care Visits Can Reduce the Risk of Potentially Avoidable Hospitalizations Among Persons With Diabetes in France,” European Journal of Public Health 30, no. 6 (2020): 1056–1061, 10.1093/EURPUB/CKAA137. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data S1: Methods used to calculate estimated prevalence of diabetes.
Table S1: Selected potentially avoidable acute complications and ICD‐10 codes.
Table S2: Characteristics of non‐preventable emergency admissions.
Table S3: Crude rates for selected complications per 1000 people with diabetes.
Table S4: Crude rates of potentially avoidable hospitalisations per 1000 people with diabetes.
Table S5: Crude rates of potentially avoidable complications by diabetes type and age group per 1000 people with diabetes.
Table S6: Age standardised rates by diabetes type and sex per 1000 people with diabetes.
Table S7: Age standardised rates by diabetes type and sex per 1000 people with diabetes—Diabetes specific events.
Table S8: Age standardised rates by diabetes type and sex per 1000 people with diabetes—Vascular events.
Table S9: Age standardised rates by diabetes type and sex per 1000 people with diabetes—Infections.
Table S10: Age standardised rates by diabetes type and sex per 1000 people with diabetes—Lower limb.
Table S11: Age standardised rates by diabetes type and sex per 1000 people with diabetes—Acute kidney failure.
Figure S1: Age standardised rates of acute complications per 1000 people with diabetes without COVID as additional diagnosis.
Figure S2: Join point analysis for all complications with and without COVID‐19 as an additional diagnosis.
Figure S3: Join point analysis for all complications excluding year 2020.
Figure S4: Join point analysis for diabetes specific hospitalisations.
Figure S5: Join point analysis for infection related hospitalisations.
Figure S6: Join point analysis for cardiovascular related hospitalisations.
Figure S7: Join point analysis for lower limb related hospitalisations.
Figure S8: Join point analysis for acute kidney failure related hospitalisations.
Figure S9: Age standardised rates of grouped conditions by diabetes and sex.
Data Availability Statement
The data used in this study are not publicly available due to restrictions under Irish and European data protection legislation, in order to protect patient privacy. The data were derived from the Hospital In‐Patient Enquiry (HIPE) system, which is managed by the Healthcare Pricing Office (HPO) on behalf of the Health Service Executive (HSE) in Ireland. Access to HIPE data may be granted to researchers with appropriate ethical approvals and who meet the criteria for access to confidential data, subject to approval by the HPO and relevant data governance processes. Requests for access to the data should be directed to the Healthcare Pricing Office. The analytical code used to support the findings of this study may be made available by the corresponding author upon reasonable request, where permitted by data governance restrictions.