Abstract
Background
Many individuals with chronic obstructive pulmonary disease (COPD) experience frequent hospitalization and readmissions, which is burdensome on the health system. This study aims to investigate factors associated with unplanned readmissions and mortality following a COPD-related hospitalization over a 12-month period in Australia, focusing on mental disorders and accounting for the acute phase of the COVID-19 pandemic.
Methods
A retrospective cohort study using linked hospitalization and mortality records identified individuals aged ≥40 years who had at least one hospital admission with a principal diagnosis of COPD between 2014 and 2020 in New South Wales, Australia. A semi-competing risk analysis was conducted to examine factors associated with unplanned readmission and mortality.
Results
Adults with a mental disorder diagnosis, specifically anxiety, had a higher risk of 12-month unplanned readmission. Individuals with anxiety and dementia also had a higher risk of mortality pre- and post-unplanned readmission. Individuals who were admitted during the acute phase of the COVID-19 pandemic period had lower risk of unplanned readmission, but higher risk of mortality without unplanned readmission.
Conclusion
Interventions aimed at reducing admissions should consider adults living with mental disorders such as anxiety or dementia to improve healthcare delivery and health outcomes for individuals living with COPD.
Keywords: chronic obstructive pulmonary disease, hospitalization, mortality, readmissions
Introduction
Chronic obstructive pulmonary disease (COPD) is a leading cause of death worldwide,1 and presents one of the highest global disease burdens.2 In the United States (US), annual healthcare costs for COPD are estimated at USD$29 billion,3 while in the United Kingdom (UK), the annual burden of COPD is estimated at £1.9 billion.4 In Australia, approximately 55% of the AUD$934 million annual cost of COPD is from treatment in Australian hospitals.5
Many individuals living with COPD experience recurrent hospitalizations. Around 20% of patients hospitalized for an acute exacerbation of COPD are readmitted within 30 days, and approximately half of these readmissions are for non-respiratory causes.6–8 In the US, COPD-related readmission costs are around 20% higher than initial hospitalization costs,8 and 1:10 of these readmissions are considered to be potentially avoidable.9
Prior studies have examined predictors of readmission and mortality following COPD-related hospitalization, including factors such as age, sex and ethnicity,10–13 comorbid conditions,8,14,15 and history of health service use.9,15–17 Mental disorders, including depression and anxiety, have been identified as highly comorbid with COPD18,19 with several studies finding mental disorders to be predictive of hospital readmissions for individuals with COPD.9,13,20–23 However, previous studies were limited by their use of health insurance-related data21,23 or data from single health providers.9,20 No studies have examined the role of mental disorders in readmissions for individuals with COPD using data from the predominantly publicly funded Australian health system.
The COVID-19 pandemic has had an impact on hospital admissions both in Australia and internationally, with studies finding a reduction in COPD admissions across the pandemic.24,25 However, 30-day mortality estimates for patients admitted with COPD are inconsistent between studies.24,26 In examining predictors of readmission for individuals with COPD, no studies have taken into account the COVID-19 period alongside other risk factors. This study aims to investigate the factors associated with unplanned readmissions and mortality following a COPD-related hospitalization over a 12-month period in Australia, focusing on mental disorders and accounting for the acute phase of the COVID-19 pandemic.
Method
Study design and study population
A retrospective cohort study using linked hospitalization and mortality records identified adults aged ≥40 years who had at least one hospital admission with a principal diagnosis of COPD between 1 January 2014 and 31 December 2020 in New South Wales (NSW), Australia. Ethical approval and a waiver of consent was obtained from the NSW Population and Health Services Research Ethics Committee (approval number 2022/ETH00861).
Data source and linkage
NSW hospitalization data includes both public and private hospitals, with dates of hospital episodes and up to 51 diagnosis and procedure codes classified using the International Statistical Classification of Diseases and Related Health Problems, Australian modification (ICD-10-AM). The hospitalization data were provided for the period 1 January 2013 to 31 December 2021 with private hospital admissions to 30 June 2021. All hospital episodes of care related to the same admission were grouped into a period of hospital care. All episodes of care provided in a public in-patient aged care facility were excluded from the count of hospitalizations.
Mortality data was obtained from death registrations (fact of death) from the NSW Registry of Birth, Deaths and Marriages for the period 1 January 2014 to 30 June 2022. The Centre for Health Record Linkage conducted probabilistic matching to link the hospitalization and mortality data, and was supplemented by manual clerical review. Logic errors (e.g. admission post-death) were identified, and all admissions related to these patients were removed.
Case definition
COPD admissions, both planned and unplanned, were identified from the hospitalization data with a principal diagnosis of COPD (ICD-10-AM: J40, J41, J42, J43, J44). The first COPD admission between 1 January 2014 and 31 December 2020 was designated as the index admission for each patient.
Identification of covariates
The Australian Statistical Geography Standard (ASGS) 2016 categories of remoteness27 were used to derive urban or rural classification, based on the individual’s residential Statistical Areas Level 2 (SA2). The Index of Relative Socio-Economic Disadvantage (IRSD) of Socio-Economic Index for Areas (SEIFA) was also assigned based on the individual’s residential SA2.28
The number of comorbidities was calculated using the Charlson Comorbidity Index29 using a 12-month lookback period with COPD and dementia excluded from the count. The number of comorbidities was categorized as nil, 1–2 and ≥3 comorbidities. Due to clinical relevance to COPD, acute myocardial infarction, congestive heart failure and renal disease were identified separately using the Charlson Comorbidity Index. Mental and behavioural disorders were identified according to ICD-10-AM Chapter V and dementia was identified by including two additional codes (ICD-10-AM: G30, G31; Appendix A). Past and current smoking status and obesity were identified using corresponding ICD-10-AM codes within the 12-month lookback period.
Frailty risk was aggregated using the validated Hospital Frailty Risk Score30 which assigns a score ranging from 0.1 and 7.1 to each of the 109 ICD-10-AM diagnosis or external cause codes based on its frailty characteristics. Dementia and depression, as conditions of interest, were excluded from the final score. The score for an individual was calculated using ICD-10-AM codes in hospitalization records in the 12-month lookback period. Frailty risk was aggregated as low (<5), intermediate (5-15), or high (>15).
Unplanned readmission and mortality
All-cause acute and emergency admissions, including both public and private hospitals’ post-index admission date of discharge, were identified as unplanned readmissions. If more than one readmission occurred within the follow-up period, the first readmission was considered the event of interest. All deaths were identified and deaths during the index admission were excluded from subsequent analysis.
Data management and analysis
Data management was conducted using SAS version 9.4 (SAS Institute, Cary NC) and statistical analysis used R version 4.2.2 with packages survival (3.5-3) SemiCompRisks (3.4).31 Number and percentages were calculated for categorical variables. For continuous variables, mean and standard deviation (SD) were calculated. The chi-square test of independence or Fisher’s exact test and Kruskal–Wallis rank sum test were used to determine differences between demographic characteristics by outcome.
All patients in the study population who were discharged alive were followed for 365 days post-discharge or to death. To understand the complex relationship between unplanned readmission and death, illness-death models were built using all-cause unplanned readmission after discharge, with death before end of follow-up as a semi-competing event.32 The Grambsch, Therneau and Fleming plot33 was used to assess the distribution of data. Time to unplanned readmission and/or death were modelled using parametric regression following the Weibull distribution with post-readmission death modelled using the Markov assumption with subject specific frailty. An indicator variable was created for individuals with >1 unplanned readmission and this indicator was included in the modelling for post-readmission death.
Variables previously identified as associated with hospital readmission and available in the data were assessed for modelling.8,15,26,34 An indicator variable was added to identify hospital episodes from the start of the first COVID-19 cases recorded in NSW (13 January 2020) to the end of the study period. Univariate Weibull time-to-event regression was used to assess factors associated with unplanned readmission; those that did not significantly contribute to risk of unplanned readmission at p <0.1 were excluded. The final multivariable cause-specific model included: age group, sex, Aboriginality (Y/N), number of Charlson comorbidities (nil, 1–2, ≥3), frailty risk (low, moderate, high), urban or rural status, past and/or current smoking status (Y/N), mental disorder diagnosis (Y/N; or breakdown in different mental disorders), renal disease (Y/N), congestive heart failure (Y/N), residence in an aged care facility (Y/N) and index admission during the COVID-19 pandemic period (Y/N). Hazard ratios (HR) were calculated, and 95% confidence intervals (CIs) reported.
Results
There were 64 850 adults aged ≥40 years who had at least one COPD admission between 1 January 2014 and 31 December 2020, and 2450 (3.8%) died during the index admission. Of the 62 400 adults discharged alive, more than half (n = 34 513, 55.3%) had at least one unplanned readmission, and 15.8% died during the 365-day follow-up period. Of those who had at least one unplanned readmission, 22.6% (n = 7811) died during the follow-up period, and over half (n = 18 678; 54.1%) had more than one unplanned readmission (Supplementary Fig. 1). The risk of unplanned readmission and death post-readmission decreased rapidly within the first month post-discharge (Fig. 1). The risk of death prior to an unplanned readmission was relatively small throughout the 12-month follow-up period.
Fig. 1.
Extended readmission-mortality model with semi-competing risks for unplanned readmission and death prior to readmission, and death after unplanned readmission in days, linked NSW hospitalization and aged care records, 1 January 2014 to 30 December 2021.
Individuals who had at least one unplanned readmission and who were alive at the end of the follow-up period were younger (mean age 73.2, SD 11.5), had a higher proportion of females (49.9%), and a higher proportion with no comorbidities (52.3%) compared with those who died following an unplanned readmission (Table 1). Individuals who had no unplanned readmissions and who were alive at the end of the follow-up period were younger (mean age 70.5, SD 11.6) and had a higher proportion of females (53.2%) compared with individuals who died pre- or post-readmission. Of those who had no readmission or death, nearly two-thirds (64.8%) had no comorbidities and 80% had a low frailty risk.
Table 1.
Characteristics at index admission of adults aged ≥40 years who had a COPD hospital admission between 1 January 2013 and 31 December 2021 by outcome at the end of 12-month follow-up in NSW
| Characteristic | No event | Readmission without death | Death prior to readmission | Death post-readmission | p-valuea |
|---|---|---|---|---|---|
| n = 25 837 | n = 26 702 | n = 2050 | n = 7811 | ||
| Age, mean (SD) | 70.5 (11.6) | 73.2 (11.5) | 79.7 (10.7) | 77.7 (10.3) | < 0.0001 |
| Age group | < 0.0001 | ||||
| 40–64 | 7601 (29.4%) | 5831 (21.8%) | 210 (10.2%) | 863 (11.0%) | |
| 65–74 | 8159 (31.6%) | 7741 (29.0%) | 380 (18.5%) | 1873 (24.0%) | |
| 75–84 | 7050 (27.3%) | 8613 (32.3%) | 663 (32.3%) | 2887 (37.0%) | |
| ≥85 | 3027 (11.7%) | 4517 (16.9%) | 797 (38.9%) | 2188 (28.0%) | |
| Sex | < 0.0001 | ||||
| Male | 12 083 (46.8%) | 13 375 (50.1%) | 1127 (55.0%) | 4537 (58.1%) | |
| Female | 13 754 (53.2%) | 13 327 (49.9%) | 923 (45.0%) | 3274 (41.9%) | |
| Index admission LOS (days) | 5.8 (21.6) | 6.8 (18.7) | 13.4 (39.4) | 9.8 (18.7) | < 0.0001 |
| ICU stay during index admission | 1042 (4.0%) | 1024 (3.8%) | 95 (4.6%) | 317 (4.1%) | 0.2532 |
| Index admission during COVID pandemic | 1868 (7.2%) | 1734 (6.5%) | 167 (8.1%) | 547 (7.0%) | 0.001 |
| Total days in hospital due to unplanned readmission | – | 25.8 (58.4) | – | 33.6 (43.7) | < 0.0001 |
| More than one unplanned readmission | – | 14 209 (53.2%) | – | 4469 (57.2%) | < 0.0001 |
| Aged care resident | 538 (2.1%) | 603 (2.3%) | 408 (19.9%) | 535 (6.8%) | < 0.0001 |
| Residential location | < 0.0001 | ||||
| Urban | 14 534 (58.5%) | 16 489 (62.5%) | 1195 (58.5%) | 4810 (61.9%) | |
| Rural | 10 307 (41.5%) | 9893 (37.5%) | 849 (41.5%) | 2957 (38.1%) | |
| Not known | 996 | 320 | 6 | 44 | |
| SEIFA IRSD quintile | 0.0004 | ||||
| 1 (most disadvantaged) | 7258 (29.2%) | 7856 (29.8%) | 556 (27.2%) | 2324 (29.9%) | |
| 2 | 7577 (30.5%) | 7685 (29.1%) | 666 (32.6%) | 2348 (30.2%) | |
| 3 | 4559 (18.4%) | 5056 (19.2%) | 348 (17.0%) | 1415 (18.2%) | |
| 4 | 2475 (10.0%) | 2701 (10.2%) | 209 (10.2%) | 813 (10.5%) | |
| 5 (least disadvantaged) | 2970 (12.0%) | 3080 (11.7%) | 265 (13.0%) | 865 (11.1%) | |
| Not known | 998 | 324 | 6 | 46 | |
| Number of comorbidities b | < 0.0001 | ||||
| nil | 16 735 (64.8%) | 13 955 (52.3%) | 782 (38.1%) | 2857 (36.6%) | |
| 1–2 | 8000 (31.0%) | 10 205 (38.2%) | 961 (46.9%) | 3575 (45.8%) | |
| ≥ 3 | 1102 (4.3%) | 2542 (9.5%) | 307 (15.0%) | 1379 (17.7%) | |
| Frailty risk | < 0.0001 | ||||
| Low | 20 681 (80.0%) | 17 992 (67.4%) | 908 (44.3%) | 3846 (49.2%) | |
| Intermediate | 4542 (17.6%) | 7312 (27.4%) | 857 (41.8%) | 3130 (40.1%) | |
| High | 614 (2.4%) | 1398 (5.2%) | 285 (13.9%) | 835 (10.7%) | |
| Smoking (past & current) | 19 469 (75.3%) | 20 736 (77.7%) | 1434 (70.1%) | 5997 (76.5%) | < 0.0001 |
| Obesity | 2291 (8.9%) | 2747 (10.3%) | 133 (6.5%) | 587 (7.5%) | < 0.0001 |
| Mental and behaviour disorders | 7291 (28.2%) | 9277 (34.7%) | 919 (44.8%) | 3242 (41.5%) | < 0.0001 |
| Dementia | 452 (1.7%) | 758 (2.8%) | 238 (11.6%) | 493 (6.3%) | < 0.0001 |
| Depression | 162 (0.6%) | 397 (1.5%) | 25 (1.2%) | 101 (1.3%) | < 0.0001 |
| Anxiety | 977 (3.8%) | 1552 (5.8%) | 137 (6.7%) | 584 (7.5%) | < 0.0001 |
| Other mental disorder | 6466 (25.0%) | 8169 (30.6%) | 710 (34.6%) | 2683 (34.3%) | < 0.0001 |
| Congestive heart failure | 2824 (10.9%) | 5024 (18.8%) | 625 (30.5%) | 2378 (30.4%) | < 0.0001 |
| Renal disease | 1292 (5.0%) | 2702 (10.1%) | 320 (15.6%) | 1382 (17.7%) | < 0.0001 |
Pearson's Chi-squared test; Kruskal-Wallis rank sum test; Wilcoxon rank sum test
excluding COPD, mental health and behaviour disorders.
Individuals who died during the follow-up period without an unplanned readmission were older (≥85 years; 38.9%), with almost one-fifth (19.9%) residing in an aged care facility. A higher proportion of individuals who died without readmission had a mental disorder (44.8%), mainly dementia (11.6%), compared with individuals with no readmission or who had a readmission without death.
Individuals aged ≥85 years (HR 1.66, 95% CI 1.59–1.73), with ≥3 comorbidities (HR 1.66, 95% CI 1.57–1.76), who had a high frailty risk (HR 1.55, 95% CI 1.46–1.65), or at least one mental disorder diagnosis (HR 1.19, 95% CI 1.16–1.22) had a higher risk of having an unplanned readmission compared to reference categories. Females (HR 0.88, 95% CI 0.86–0.90), individuals living in a rural area (HR 0.94, 95% CI 0.92–0.97), those living in an aged care facility (HR 0.85, 95% CI 0.79–0.91), and those with an index admission during the COVID-19 pandemic (HR 0.93, 95% CI 0.88–0.97) all had a lower risk of having an unplanned readmission compared to reference categories (Table 2).
Table 2.
Semi-competing risk analysis of common factors associated with unplanned hospital readmission and mortality up to 12-month follow-up, linked NSW hospitalization and death records, 1 January 2013 to 31 December 2021
| Readmission | Death prior to readmission | Death post-readmission | ||||
|---|---|---|---|---|---|---|
| Adjusted HR | 95% CI | Adjusted HR | 95% CI | Adjusted HR | 95% CI | |
| Sex | ||||||
| Male | 1 | – | 1 | – | 1 | – |
| Female | 0.88 | (0.86, 0.90) | 0.73 | (0.66, 0.80) | 0.76 | (0.73, 0.80) |
| Age group | ||||||
| 40–64 | 1 | – | 1 | – | 1 | – |
| 65–74 | 1.23 | (1.19, 1.28) | 1.39 | (1.17, 1.65) | 1.53 | (1.40, 1.66) |
| 75–84 | 1.49 | (1.43, 1.54) | 2.13 | (1.81, 2.50) | 1.98 | (1.82, 2.15) |
| ≥85 | 1.66 | (1.59, 1.73) | 3.87 | (3.27, 4.58) | 2.7 | (2.47, 2.96) |
| Number of comorbidities a | ||||||
| nil | 1 | – | 1 | – | 1 | – |
| 1–2 | 1.27 | (1.23, 1.31) | 1.55 | (1.39, 1.74) | 1.41 | (1.33, 1.5) |
| ≥3 | 1.66 | (1.57, 1.76) | 1.93 | (1.61, 2.31) | 1.79 | (1.63, 1.97) |
| Frailty risk | ||||||
| Low | 1 | – | 1 | – | 1 | – |
| Intermediate | 1.41 | (1.37, 1.46) | 2.02 | (1.81, 2.24) | 1.5 | (1.42, 1.59) |
| High | 1.55 | (1.46, 1.65) | 2.57 | (2.19, 3.02) | 1.61 | (1.47, 1.77) |
| Residential location | ||||||
| Urban | 1 | – | 1 | – | 1 | – |
| Rural | 0.94 | (0.92, 0.97) | 1.25 | (1.14, 1.37) | 1.16 | (1.11, 1.22) |
| Mental disorder diagnosis | ||||||
| No | 1 | – | 1 | – | 1 | – |
| Yes | 1.19 | (1.16, 1.22) | 1.43 | (1.30, 1.58) | 1.21 | (1.15, 1.27) |
| Aged care resident | ||||||
| No | 1 | – | 1 | – | 1 | – |
| Yes | 0.85 | (0.79, 0.91) | 3.84 | (3.39, 4.35) | 2.05 | (1.85, 2.28) |
| Index admission during COVID-19 pandemic | ||||||
| No | 1 | – | 1 | – | 1 | – |
| Yes | 0.93 | (0.88, 0.97) | 1.18 | (1.00, 1.39) | 1.05 | (0.96, 1.16) |
Abbreviation: Adjusted HR, adjusted hazard ratio.
Note: model also adjusted for Aboriginality, past and current smoking status, congestive heart failure, renal disease diagnosis during lookback periods. Death post-readmission model component also adjusted for multiple unplanned readmissions.
excluding COPD, mental health and behaviour disorders.
Being in an older age group, having a higher number of comorbidities and increased frailty were all associated with an increased risk of death either pre- or post- unplanned readmission compared with reference groups. Individuals who lived in a rural area also had higher risk of death both pre- (HR 1.25, 95% CI 1.14–1.37) and post- (HR 1.16, 95% CI 1.11–1.22) unplanned readmission compared with individuals in urban areas. Individuals who had their index admission during the acute phase of the COVID-19 pandemic had increased risk of death without having an unplanned readmission (HR 1.18, 95% CI 1.00–1.39), but no statistically significant difference in the risk of death post-readmission, compared to individuals who had index admission prior to COVID-19 pandemic.
A mental disorder diagnosis was associated with having an unplanned readmission (HR 1.19, 95% CI 1.16–1.22), death pre- (HR 1.43, 95% CI 1.30–1.58) and post- (HR 1.21, 95% CI 1.15–1.27) unplanned readmission compared with adults who did not have a mental disorder diagnosis. When modelled on different mental disorders, individuals with an anxiety disorder had 1.33 times the risk (95% CI 1.25–1.41) of having an unplanned readmission, 1.33 times the risk (95% CI 1.09–1.62) of death pre- unplanned readmission, and 1.38 times the risk (95% CI 1.25–1.53) of death post-readmission compared to adults without an anxiety disorder diagnosis. Having a dementia diagnosis was associated with higher risk of death both pre- (HR 1.70, 95% CI 1.46–1.98) and post- (HR 1.35, 95% CI 1.21–1.50) unplanned readmission (Table 3).
Table 3.
Semi-competing risk analysis of common factors and separate mental disorders diagnosis associated with unplanned hospital readmission and mortality up to 12-month follow-up, linked NSW hospitalization and death records, 1 January 2013 to 31 December 2021
| Readmission | Death prior to readmission | Death post-readmission | ||||
|---|---|---|---|---|---|---|
| Adjusted HR | 95% CI | Adjusted HR | 95% CI | Adjusted HR | 95% CI | |
| Sex | ||||||
| Male | 1 | – | 1 | – | 1 | – |
| Female | 0.88 | (0.85, 0.90) | 0.73 | (0.67, 0.80) | 0.76 | (0.72, 0.80) |
| Age group | ||||||
| 40–64 | 1 | – | 1 | – | 1 | – |
| 65–74 | 1.24 | (1.19, 1.28) | 1.37 | (1.16, 1.63) | 1.52 | (1.39, 1.65) |
| 75–84 | 1.5 | (1.44, 1.56) | 2.07 | (1.75, 2.43) | 1.95 | (1.79, 2.12) |
| ≥85 | 1.68 | (1.61, 1.75) | 3.71 | (3.13, 4.40) | 2.65 | (2.42, 2.90) |
| Comorbidity count a | ||||||
| nil | 1 | – | 1 | – | 1 | – |
| 1–2 | 1.27 | (1.24, 1.31) | 1.55 | (1.39, 1.74) | 1.41 | (1.33, 1.50) |
| ≥3 | 1.67 | (1.58, 1.77) | 1.96 | (1.63, 2.35) | 1.79 | (1.63, 1.97) |
| Frailty risk | ||||||
| Low | 1 | – | 1 | – | 1 | – |
| Intermediate | 1.40 | (1.36, 1.45) | 2.01 | (1.81, 2.24) | 1.50 | (1.42, 1.59) |
| High | 1.52 | (1.43, 1.61) | 2.53 | (2.15, 2.98) | 1.61 | (1.46, 1.76) |
| Residential location | ||||||
| Urban | 1 | – | 1 | – | 1 | – |
| Rural | 0.94 | (0.92, 0.97) | 1.25 | (1.14, 1.37) | 1.16 | (1.11, 1.22) |
| Anxiety diagnosis | ||||||
| No | 1 | – | 1 | – | 1 | – |
| Yes | 1.33 | (1.25, 1.41) | 1.33 | (1.09, 1.62) | 1.38 | (1.25, 1.53) |
| Depression diagnosis | ||||||
| No | 1 | – | 1 | – | 1 | – |
| Yes | 1.33 | (1.17, 1.5) | 0.91 | (0.59, 1.41) | 0.65 | (0.52, 0.82) |
| Dementia diagnosis | ||||||
| No | 1 | – | 1 | – | 1 | – |
| Yes | 1.09 | (1.02, 1.17) | 1.70 | (1.46, 1.98) | 1.35 | (1.21, 1.5) |
| Other mental and behaviour disorder diagnosis | ||||||
| No | 1 | – | 1 | – | 1 | – |
| Yes | 1.14 | (1.11, 1.18) | 1.22 | (1.10, 1.35) | 1.09 | (1.04, 1.16) |
| Aged care resident | ||||||
| No | 1 | – | 1 | – | 1 | – |
| Yes | 0.85 | (0.79, 0.91) | 3.69 | (3.25, 4.19) | 2.03 | (1.83, 2.25) |
| Index admission during COVID-19 pandemic | ||||||
| No | 1 | – | 1 | – | 1 | – |
| Yes | 0.93 | (0.89, 0.98) | 1.19 | (1.01, 1.4) | 1.06 | (0.96, 1.16) |
Abbreviation: Adjusted HR, adjusted hazard ratio.
Note: model also adjusted for Aboriginality, past and current smoking status, congestive heart failure, renal disease diagnosis during lookback periods. Death post-readmission model component also adjusted for multiple unplanned readmissions.
excluding COPD, mental health and behaviour disorders.
Discussion
Main findings of this study
This study examined the factors associated with unplanned readmissions and mortality following a COPD-related hospitalization over a 12-month period in Australia, focusing on mental disorders and accounting for the acute phase of the COVID-19 pandemic. Being of older age, being female, having a higher number of comorbidities, having higher frailty risk, a mental disorder diagnosis, living in rural areas, and having an index admission during the COVID-19 period were found to be predictive of an unplanned hospital readmission.
What is already known on this topic & what this study adds
Individuals with a mental disorder, specifically anxiety, had a higher risk of unplanned readmission compared with individuals without these diagnoses. Mental disorders, such as anxiety and depression, have previously been predictive of hospital admissions of individuals with COPD.9,23 The effects of anxiety on an individual with COPD are bidirectional,35 and can lead to a high symptom burden for individuals. In relation to readmissions, it is possible that a COPD-related breathing difficulty may be misattributed as a symptom of anxiety, or vice versa, which could lead to heightened arousal and an increased need for readmission.23
A higher risk of death both pre- and post- unplanned readmission was also found in this study for COPD patients with a mental disorder diagnosis, specifically anxiety or dementia. A prior meta-analysis found comorbid depression and anxiety increased the risk of COPD exacerbations and mortality, as poor mental health can hinder adherence to treatment plans and COPD prognosis.35 Individuals with mental disorders, especially those with dementia, may also have limited social support and reduced capacity to self-manage COPD symptoms.19,23,36 These factors may contribute to the higher risk of death pre- and post-admission found for COPD patients with mental disorders in this study.
In this study, individuals living in rural areas had a lower risk of unplanned readmission, and higher risk of death, both pre- and post-unplanned readmissions. A previous US study identified that adjusting for travel time, patients in rural areas had fewer COPD-related hospitalizations,37 highlighting the importance of service accessibility. In this study, individuals with COPD in rural areas may avoid hospital due to distance and may have only presented to hospital for very severe, or late-stage symptoms, reflected in the observed mortality. However, a US study found that rural living was not associated with mortality over the long-term for COPD patients.37 These differing findings to the current study may be attributed to different study populations, where the US study only included US veterans aged ≥65 years, and employed different analysis methods37 which may have influenced risk estimates.
Individuals with COPD who had their index admission during the COVID-19 pandemic had lower risk of unplanned readmission, but higher risk of death without readmission in this study. Prior literature found that hospital admissions decreased during this period,24,25 indicating that concerns about infection or overloading the health system38 may have increased the threshold for healthcare utilization. However, the lower risk of readmission and higher risk of death without readmission highlight that potentially, individuals may have had severe symptoms but were not presenting to hospital. Alternatively, individuals with COPD admissions during this period may have been differentially affected by mental or physical health challenges,25 thus worsening the burden of COPD and contributing to the observed mortality.
Limitations of this study
The strengths of this study include that it used population-level linked health data, a 12-month lookback period, a 12-month follow-up period, and investigated mental disorders in an Australian context. However, the prevalence of mental disorders is lower in this study than previous cohorts,18 as only diagnoses related to episodes of care were recorded in the administrative data, compared with other identification methods.39 The lack of clinical measures in the dataset limited the study, as disease severity was unable to be adjusted for. Cause of death was also not available for all deaths in this study, which would broaden interpretation of all-cause versus COPD-related mortality. The interpretation of the effect of COVID-19 was also limited by its inclusion in the model only as an indicator variable.
Conclusion
The factors identified in this study, particularly mental disorders and rurality, could inform the development of a risk stratification algorithm to predict readmission or mortality outcomes for COPD patients.8,13,40 COPD-related hospitalization could then be mitigated through interventions targeted at high-risk individuals. Factors such as anxiety are treatable through non-acute healthcare services,41 which reduce the burden on hospitals and related healthcare expenditures, and improve patient outcomes for individuals living with COPD.
Supplementary Material
Acknowledgements
The authors thank the Centre for Health Record Linkage (CHeReL) for conducting the data linkage, and the NSW Ministry of Health for providing access to population health data. This work was completed while Zhisheng Sa was employed as a trainee on the NSW Biostatistics Training Program funded by the NSW Ministry of Health. Zhisheng undertook this work while based at the Australian Institute of Health Innovation, Macquarie University.
Appendix A
ICD-10-AM codes used for identifying comorbidity and conditions not included in Charlson Comorbidity Index:
All mental disorders: F00-F99, G30, G31.
Anxiety disorder: F40-F48.
Dementia: F00-F03, F05.1, G30, G31.
Depressive disorder: F20.4, F31.3, F31.4, F31.5, F32, F33, F34.1, F41.2, F43.2.
Other mental disorders: all mental disorders code excluding anxiety, depressive disorders and dementia code.
Obesity: E66, U78.1.
Smoking: F17, Z72.0, Z71.6, Z86.43.
Shalini Wijekulasuriya, Research Assistant
Zhisheng Sa, Biostatistical Trainee
Tim Badgery-Parker, PhD, Research Fellow
Janet C. Long, PhD, Senior Research Fellow
Jeffrey Braithwaite, PhD, Professor
David G. Chapman, PhD, Clinician-Researcher
Jean-Frédéric Levesque, PhD, Deputy Secretary, Chief Executive and Adjunct Professor
Diane E. Watson, PhD, Chief Executive
Johanna I. Westbrook, PhD, Professor
Rebecca Mitchell, PhD, Professor
Contributor Information
Shalini Wijekulasuriya, Australian Institute of Health Innovation, Macquarie University, Sydney, NSW, 2109, Australia.
Zhisheng Sa, Australian Institute of Health Innovation, Macquarie University, Sydney, NSW, 2109, Australia; NSW Biostatistics Training Program, NSW Ministry of Health, Sydney, NSW, 2065, Australia.
Tim Badgery-Parker, Australian Institute of Health Innovation, Macquarie University, Sydney, NSW, 2109, Australia.
Janet C Long, Australian Institute of Health Innovation, Macquarie University, Sydney, NSW, 2109, Australia.
Jeffrey Braithwaite, Australian Institute of Health Innovation, Macquarie University, Sydney, NSW, 2109, Australia.
David G Chapman, Respiratory Investigation Unit, Royal North Shore Hospital, Sydney, NSW, 2065, Australia; Woolcock Institute of Medical Research, Macquarie University, Sydney, NSW, 2109, Australia; School of Life Sciences, University of Technology Sydney, Ultimo, NSW, 2007, Australia.
Jean-Frédéric Levesque, Agency for Clinical Innovation, Sydney, NSW, 2065, Australia; Centre for Primary Health Care and Equity, University of New South Wales, Sydney, NSW, 2052, Australia.
Diane E Watson, Bureau of Health Information, Sydney, NSW, 2065, Australia.
Johanna I Westbrook, Australian Institute of Health Innovation, Macquarie University, Sydney, NSW, 2109, Australia.
Rebecca Mitchell, Australian Institute of Health Innovation, Macquarie University, Sydney, NSW, 2109, Australia.
Declaration of interest
The authors have no competing interests to declare that are relevant to the content of this article.
Funding
This study was supported by the Medical Research Future Fund (MRFF), Keeping Australians Out of Hospital (APP1178554).
Authors' contributions
SW and ZS wrote the first draft of the manuscript. ZS conducted the data analysis with assistance from RM and TBP. All authors (i.e., SW, ZS, TBP, JL, JB, DC, JFL, DW, JW, RM) edited and approved the final manuscript.
Ethics approval
Ethical approval and a waiver of consent was obtained from the NSW Population and Health Services Research Ethics Committee (approval number 2022/ETH00861) in view of the retrospective nature of the study.
Data availability
The data that support the findings of this study are available from the NSW Ministry of Health. Restrictions apply to the availability of these data, which were used under licence for the current study, so are not publicly available.
References
- 1. World Health Organization . Chronic obstructive pulmonary disease (COPD). 2023. Geneva, Switzerland: World Health Organization. Available from: https://www.who.int/news-room/fact-sheets/detail/chronic-obstructive-pulmonary-disease-(copd).
- 2. The Global Health Observatory . Global Health Estimates: Life Expectancy and leading causes of death and disability. 2023. Geneva, Switzerland: World Health Organization. Available from: https://www.who.int/data/gho/data/themes/mortality-and-global-health-estimates.
- 3. American Lung Association . COPD Trends Brief: Burden. 2023. Chicago, Illinois: American Lung Association. Available from: https://www.lung.org/research/trends-in-lung-disease/copd-trends-brief/copd-burden.
- 4. NHS England . Respiratory Disease. 2023. Leeds, England: NHS England. Available from: https://www.england.nhs.uk/ourwork/clinical-policy/respiratory-disease/.
- 5. Australian Institute of Health and Welfare . COPD. Canberra, Australia: AIHW, 2023.
- 6. Buhr RG, Jackson NJ, Dubinett SM. et al. Factors associated with differential readmission diagnoses following acute exacerbations of chronic obstructive pulmonary disease. J Hosp Med 2020;15:219–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Shah T, Churpek MM, Coca Perraillon M. et al. Understanding why patients with COPD get readmitted: a large national study to delineate the Medicare population for the readmissions penalty expansion. Chest 2015;147:1219–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Jacobs DM, Noyes K, Zhao J. et al. Early hospital readmissions after an acute exacerbation of chronic obstructive pulmonary disease in the Nationwide readmissions database. Ann Am Thorac Soc 2018;15:837–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Shams I, Ajorlou S, Yang K. A predictive analytics approach to reducing 30-day avoidable readmissions among patients with heart failure, acute myocardial infarction, pneumonia, or COPD. Health Care Manag Sci 2015;18:19–34. [DOI] [PubMed] [Google Scholar]
- 10. Cerezo Lajas A, Gutiérrez González E, Llorente Parrado C. et al. Readmission due to exacerbation of COPD: associated factors. Lung 2018;196:185–93. [DOI] [PubMed] [Google Scholar]
- 11. Chan FW, Wong FY, Yam CH. et al. Risk factors of hospitalization and readmission of patients with COPD in Hong Kong population: analysis of hospital admission records. BMC Health Serv Res 2011;11:186. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Hahn EJ, Rayens MK, Adkins S. et al. Fewer hospitalizations for chronic obstructive pulmonary disease in communities with smoke-free public policies. Am J Public Health 2014;104:1059–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Lau CS, Siracuse BL, Chamberlain RS. Readmission after COPD exacerbation scale: determining 30-day readmission risk for COPD patients. Int J Chron Obstruct Pulmon Dis 2017;12:1891–902. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Baker CL, Zou KH, Su J. Risk assessment of readmissions following an initial COPD-related hospitalization. Int J Chron Obstruct Pulmon Dis 2013;8:551–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Roberts MH, Clerisme-Beaty E, Kozma CM. et al. A retrospective analysis to identify predictors of COPD-related rehospitalization. BMC Pulm Med 2016;16:68. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Nguyen HQ, Chu L, Amy Liu IL. et al. Associations between physical activity and 30-day readmission risk in chronic obstructive pulmonary disease. Ann Am Thorac Soc 2014;11:695–705. [DOI] [PubMed] [Google Scholar]
- 17. Sharma G, Kuo YF, Freeman JL. et al. Outpatient follow-up visit and 30-day emergency department visit and readmission in patients hospitalized for chronic obstructive pulmonary disease. Arch Intern Med 2010;170:1664–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Yohannes AM, Willgoss TG, Baldwin RC. et al. Depression and anxiety in chronic heart failure and chronic obstructive pulmonary disease: prevalence, relevance, clinical implications and management principles. Int J Geriatr Psychiatry 2010;25:1209–21. [DOI] [PubMed] [Google Scholar]
- 19. Mikkelsen RL, Middelboe T, Pisinger C. et al. Anxiety and depression in patients with chronic obstructive pulmonary disease (COPD). A review Nord J Psychiatry 2004;58:65–70. [DOI] [PubMed] [Google Scholar]
- 20. Abrams TE, Vaughan-Sarrazin M, Vander Weg MW. Acute exacerbations of chronic obstructive pulmonary disease and the effect of existing psychiatric comorbidity on subsequent mortality. Psychosomatics 2011;52:441–9. [DOI] [PubMed] [Google Scholar]
- 21. Sharif R, Parekh TM, Pierson KS. et al. Predictors of early readmission among patients 40 to 64 years of age hospitalized for chronic obstructive pulmonary disease. Ann Am Thorac Soc 2014;11:685–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Simmering JE, Polgreen LA, Comellas AP. et al. Identifying patients with COPD at high risk of readmission. Chronic Obstr Pulm Dis 2016;3:729–38. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Singh G, Zhang W, Kuo YF. et al. Association of psychological disorders with 30-day readmission rates in patients with COPD. Chest 2016;149:905–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Bureau of Health Information . New South Wales and the COVID-19 Pandemic from 2020 to 2022. Sydney (NSW): BHI, 2023. [Google Scholar]
- 25. Alqahtani JS, Oyelade T, Aldhahir AM. et al. Reduction in hospitalised COPD exacerbations during COVID-19: a systematic review and meta-analysis. PloS One 2021;16:e0255659, 1-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Lawless M, Burgess M, Bourke S. Impact of COVID-19 on hospital admissions for COPD exacerbation: lessons for future care. Medicina (Kaunas) 2022;58:66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Australian Bureau of Statistics . Australian Statistical Geography Standard (ASGS): Volume 5 - Remoteness Structure, July 2016. 2018. Canberra, Australia: Australian Bureau of Statistics. Available from: https://www.abs.gov.au/ausstats/abs@.nsf/Latestproducts/1270.0.55.005Main%20Features15July%202016?opendocument&tabname=Summary&prodno=1270.0.55.005&issue=July%202016&num=&view.
- 28. Australian Bureau of Statistics . Census of Population and Housing: Socio-Economic Indexes for Areas (SEIFA), Australia: 2016. 2018. Canberra, Australia: Australian Bureau of Statistics. Available from: https://www.abs.gov.au/AUSSTATS/abs@.nsf/Lookup/2033.0.55.001Main+Features12016?OpenDocument. [Google Scholar]
- 29. Sundararajan V, Henderson T, Perry C. et al. New ICD-10 version of the Charlson comorbidity index predicted in-hospital mortality. J Clin Epidemiol 2004;57:1288–94. [DOI] [PubMed] [Google Scholar]
- 30. Gilbert T, Neuburger J, Kraindler J. et al. Development and validation of a hospital frailty risk score focusing on older people in acute care settings using electronic hospital records: an observational study. Lancet 2018;391:1775–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Alvares D, Haneuse S, Lee C. et al. SemiCompRisks: an R package for the analysis of independent and cluster-correlated semi-competing risks data. R j 2019;11:376–400. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Lee KH, Haneuse S, Schrag D. et al. Bayesian semi-parametric analysis of semi-competing risks data: investigating hospital readmission after a pancreatic cancer diagnosis. J R Stat Soc Ser C Appl Stat 2015;64:253–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Grambsch PM, Therneau TM, Fleming TR. Diagnostic plots to reveal functional form for covariates in multiplicative intensity models. Biometrics 1995;51:1469–82. [PubMed] [Google Scholar]
- 34. McGhan R, Radcliff T, Fish R. et al. Predictors of rehospitalization and death after a severe exacerbation of COPD. Chest 2007;132:1748–55. [DOI] [PubMed] [Google Scholar]
- 35. Atlantis E, Fahey P, Cochrane B. et al. Bidirectional associations between clinically relevant depression or anxiety and COPD: a systematic review and meta-analysis. Chest 2013;144:766–77. [DOI] [PubMed] [Google Scholar]
- 36. Baird C, Lovell J, Johnson M. et al. The impact of cognitive impairment on self-management in chronic obstructive pulmonary disease: a systematic review. Respir Med 2017;129:130–9. [DOI] [PubMed] [Google Scholar]
- 37. Fortis S, Gao Y, Baldomero AK. et al. Association of rural living with COPD-related hospitalizations and deaths in US veterans. Sci Rep 2023;13:7887. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Imlach F, McKinlay E, Kennedy J. et al. Seeking healthcare during lockdown: challenges, opportunities and lessons for the future. Int J Health Policy Manag 2022;11:1316–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Edwards J, Thind A, Stranges S. et al. Concordance between health administrative data and survey-derived diagnoses for mood and anxiety disorders. Acta Psychiatr Scand 2020;141:385–95. [DOI] [PubMed] [Google Scholar]
- 40. van Walraven C, Dhalla IA, Bell C. et al. Derivation and validation of an index to predict early death or unplanned readmission after discharge from hospital to the community. CMAJ 2010;182:551–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Prieto-Centurion V, Markos MA, Ramey NI. et al. Interventions to reduce rehospitalizations after chronic obstructive pulmonary disease exacerbations. A systematic review. Ann Am Thorac Soc 2014;11:417–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data that support the findings of this study are available from the NSW Ministry of Health. Restrictions apply to the availability of these data, which were used under licence for the current study, so are not publicly available.