Reasons for readmission after hospital discharge in patients with chronic diseases-Information from an international dataset.

BACKGROUND: Chronic diseases are increasingly prevalent in Western countries. Once hospitalised, the chance for another hospitalisation increases sharply with large impact on well-being of patients and costs. The pattern of readmissions is very complex, but poorly understood for multiple chronic diseases.
METHODS: This cohort study of administrative discharge data between 2009-2014 from 21 tertiary hospitals (eight USA, five UK, four Australia, four continental Europe) investigated rates and reasons of readmissions to the same hospital within 30 days after unplanned admission with one of the following chronic conditions; heart failure; atrial fibrillation; myocardial infarction; hypertension; stroke; chronic obstructive pulmonary disease (COPD); bacterial pneumonia; diabetes mellitus; chronic renal disease; anaemia; arthritis and other cardiovascular disease. Proportions of readmissions with similar versus different diseases were analysed.
RESULTS: Of 4,901,584 admissions, 866,502 (17.7%) were due to the 12 chronic conditions. In-hospital, 43,573 (5.0%) patients died, leaving 822,929 for readmission analysis. Of those, 87,452 (10.6%) had an emergency 30-day readmission, rates ranged from 2.8% for arthritis to 18.4% for COPD. One third were readmitted with the same condition, ranging from 53% for anaemia to 11% for arthritis. Reasons for readmission were due to another chronic condition in 10% to 35% of the cases, leaving 30% to 70% due to reasons other than the original 12 conditions (most commonly, treatment related complications and infections). The chance of being readmitted with the same cause was lower in the USA, for female patients, with increasing age, more co-morbidities, during study period and with longer initial length of stay.
CONCLUSION: Readmission in chronic conditions is very common and often caused by diseases other than the index hospitalisation. Interventions to reduce readmissions should therefore focus not only on the primary condition but on a holistic consideration of all the patient's comorbidities.

Introduction

Chronic diseases are increasingly prevalent in Western countries due to population aging and better treatment of underlying conditions. Once hospitalised, the chance for another hospitalisation increases sharply. The wide variation in readmission rates suggests that a significant proportion is avoidable [1]. The US Hospital Readmissions Reduction Program requires Centers for Medicare and Medicaid Services to reduce payments to hospitals with excess readmissions [2].

Readmissions pose a significant burden to patients and healthcare systems as they create significant mortality and morbidity [3]. The latter is not only important regarding well-being, but also has an important economic impact in form of costly hospitalisations [4]. Therefore, one important aim in treating chronic diseases is to prevent readmissions. The high risk of readmission has been shown for different chronic diseases [5-7] and after diverse interventions such as hip replacement, hip fracture, otolaryngology, bypass or general surgery [8-12]. The reasons for readmission after an intervention are often not primarily related to the intervention itself, but to the underlying comorbid conditions. Thus, chronic diseases may play an important role in readmission risk, independently of the reason for the initial hospitalisation [13].

Chronic diseases usually do not occur in isolation. Most patients with chronic disease have multiple diseases [14]. They may influence each other, and treatment for one disease may adversely impact the other. Hospital quality also impacts readmission rates [15]. Therefore, the pattern of readmissions may be very complex. Because of the interactions between disease and treatment, knowing patterns of readmissions related to different chronic diseases may improve the understanding of this important problem and reveal new treatment modalities for these patients. Unfortunately, most previous studies focused on single disease presentation at index hospitalisation and did not investigate the interplay between different diseases [16].

Therefore, we aimed to investigate reasons for readmission in patients with an index hospitalisation for multiple chronic conditions and three acute diseases usually related to an underlying chronic condition. We used data from an international dataset created from hospitals around the world through the Global Comparators project [17]. Our hypothesis was that reasons for readmission are often different from the index hospitalisation, irrespective of the underlying condition, and that this complexity increases with age and comorbidities. We also wanted to determine if the result changed over time, to what extent the results were influenced by in-hospital mortality, and if there were country specific differences.

Methods

We collected electronic inpatient records from administrative discharge data provided by each of the 21 participating hospitals (eight from USA, five from UK, four from Australia, one each from Belgium, Denmark, Italy and the Netherlands; the latter four are combined as European centres; S1 Table), integrated into a uniform dataset as described elsewhere [17]. Institutional regulatory boards waived the requirement of Informed consent by patients and data were fully anonymised for the purpose of this analysis. For the present analysis, we used data from six complete calendar years, 2009–2014. All hospitals were major teaching hospitals. We excluded data before (i.e. 2007–2008) and after (i.e. 2015 and 2016) this period because many centres did not provide data for those periods to avoid selection bias. Between 2011–2013, complete data are available in all centres; in the two years before data from four hospitals (three in USA, one European) and in 2014 data from two UK hospitals were missing. Some hospitals reported in a previous publication [18] were excluded due to data quality issues, some missing codes and linkage issues from one year to the next.

We included records containing information on age, sex, country, principal diagnosis code (International Classification of Diseases, ICD-9 or ICD-10 [ICD-10_CM, ICD-10 AM]), Clinical Classification Software (CCS) group (Agency for Healthcare Quality and Research’s CCS groups (AHRQ CCS)) as described previously, [17] admission date, discharge date (date of death if admission ended in death) and in-hospital death. Each record was assigned a comorbidity score according to a modified version of the Elixhauser index [19], which is based on 32 conditions identified by secondary diagnoses codes [20]. Admissions were assigned to one of 259 diagnostic groups based on the primary diagnosis field (AHRQ CCS). Both bespoke chronic diseases (ICD-classification) and CCS codes were used to define the following chronic conditions as index hospitalisations: heart failure (HF); atrial fibrillation (AF); myocardial infarction (AMI); hypertension; stroke; chronic obstructive pulmonary disease (COPD); bacterial pneumonia; diabetes mellitus; chronic renal disease; anaemia; arthritis and other cardiovascular disease (see S2 Table). We included three acute conditions, i.e. stroke, AMI and bacterial pneumonia; AMI because it represents the most important event of an important underlying chronic condition, i.e. coronary artery disease, and pneumonia because it is often related to underlying chronic lung diseases, particularly COPD.

We excluded records corresponding to planned day-cases. Due to the difficulty in some countries in distinguishing patients admitted for observation only from those admitted as inpatients, we also excluded short-term emergency admissions with length of stay (LOS) less than two nights and no surgery.

Rates and reasons for readmissions within 30 days were analysed, focusing on the comparison between similar and different reasons for readmission compared with the index hospitalisation. Descriptive comparisons were made for each of the above-mentioned diseases. For a better illustration, Sankey diagrams are used to show the readmission patterns for each of these conditions [21]. Readmissions could only be identified when readmission was to the same hospital. For the analysis investigating readmission rates, in-hospital deaths were not considered. We allowed for a three-month window at the end of the time-period to ensure capturing the majority of readmissions for each hospital. For consistency across countries, readmissions with discharge after a three-month window were not counted, which accounted for <0.5% of the cases.

We further compared the readmission distributions for each of the 12 diseases, given by the proportion of readmissions to each of the 12 conditions or any CCS group. We calculated the Hellinger distance [22] between each pair, to identify clusters within the readmission patterns of the 12 diseases. By definition, the Hellinger distance is between zero and one, where zero means that patterns are identical, the larger the distance the more dissimilar the patterns.

To investigate the proportion of readmissions with similar versus different diseases we performed logistic regression with the outcome: readmission with the same versus any other condition. The model was adjusted for age, sex, year of discharge, comorbidity score, country, chronic condition and LOS. We applied a square root transformation to LOS as data are highly skewed to the right. Variables were considered significant when p<0.05. All data analyses were done using R, v. 3.4.2 [23].

Results

Out of 4,901,586 admissions, 866,502 (17.7%) were due to the 12 bespoke conditions as primary diagnosis (UK n = 207,748, 15.4%; continental Europe n = 187,182, 17.7%; Australia n = 161,171, 18.0%; USA n = 310,401, 19.4% USA). The patient characteristics are shown in Table 1. Patients in the USA were younger compared with the other countries, had a shorter LOS and more co-morbidities recorded. There were significant differences in baseline characteristics between the different conditions. Average LOS varied by a factor of almost three between the different conditions and differed between countries (Table 1). Table 2 shows the most common CCS groups / chronic conditions by number of admissions, indicating that the selected conditions are among the most important reasons for hospitalisation. Fig 1 provides an overview of the reasons of admission in the different countries over time.

Table 1

Patient characteristics.

	Average age	Average number of comorbidities	% male	Average length of stay (LOS)	Median LOS
Overall	65.4	2.29	56.6	7.5	4
Country
Australia	68.4	1.81	58.1	8.5	4
Continental Europe	66.6	1.38	57.8	7.7	5
England	66.3	1.98	58.6	9.3	5
USA	62.5	3.30	53.6	5.8	4
Chronic condition
AF	65.6	1.80	57.0	4.5	3
AMI	66.9	2.26	69.9	5.7	3
Anemia	52.4	1.68	45.3	6.2	4
Arthritis	63.9	1.46	42.1	4.3	3
COPD	68.6	1.77	51.0	7.6	5
Diabetes	52.4	2.20	56.5	7.8	4
Heart failure	70.6	3.57	55.5	8.9	6
Hypertension	62.4	1.49	43.3	4.9	3
Other cardiovascular	67.7	2.10	60.9	8.5	4
Pneumonia	64.2	2.31	53.3	8.9	6
Renal failure	63.1	3.06	56.7	8.5	5
Stroke	68.5	2.53	50.3	12.7	6

Abbreviations: AF atrial fibrillation; AMI acute myocardial infarction; COPD chronic obstructive pulmonary disease

Table 2

Most common CCS groups and chronic conditions as primary cause of initial hospitalization.

Condition group	Volume	Percentage
CCS group Liveborn	231377	4.72%
Chronic condition AMI	175437	3.58%
CCS group Complication of device, implant or graft	115533	2.36%
Chronic condition Pneumonia	97496	1.99%
CCS group Other complications of birth, puerperium affecting management of mother	90551	1.85%
Chronic condition Stroke	86189	1.76%
Chronic condition Heart failure	84963	1.73%
Chronic condition AF	82723	1.69%
CCS group Complications of surgical procedures or medical care	80353	1.64%
CCS group Spondylosis, intervertebral disc disorders, other back problems	75086	1.53%
Chronic condition Other cardiovascular disease	72414	1.48%
CCS group Other nervous system disorders	66199	1.35%
CCS group Biliary tract disease	65289	1.33%
Chronic condition COPD	63450	1.29%
CCS group Secondary malignancies	61933	1.26%
CCS group Urinary tract infections	61488	1.25%
CCS group Residual codes, unclassified	60878	1.24%
Chronic condition Arthritis	59969	1.22%
CCS group Septicemia (except in labour)	59506	1.21%
Chronic condition Chronic renal failure	57209	1.17%
CCS group Skin and subcutaneous tissue infections	56959	1.16%
CCS group Trauma to perineum and vulva	52616	1.07%
CCS group Normal pregnancy and/or delivery	51040	1.04%
CCS group Maintenance chemotherapy, radiotherapy	50700	1.03%
CCS group Fracture of upper limb	49437	1.01%
CCS group Foetal distress and abnormal forces of labour	48947	1.00%
CCS group Rehabilitation care, fitting of prostheses, and adjustment of devices	46808	.95%
CCS group Fracture of lower limb	46628	.95%
CCS group Other and unspecified benign neoplasm	45375	.93%
CCS group Epilepsy, convulsions	44791	.91%
CCS group Osteoarthritis	44689	.91%
CCS group Abdominal hernia	44214	.90%
CCS group Other connective tissue disease	44197	.90%
CCS group Other complications of pregnancy	44190	.90%
Chronic condition Diabetes	41136	.84%
Chronic condition Anemia	38163	.78%
CCS group Heart valve disorders	38089	.78%
CCS group Affective disorders	37311	.76%
CCS group Fracture of neck of femur (hip)	37177	.76%
CCS group Other fractures	36657	.75%
CCS group Appendicitis and other appendiceal conditions	36167	.74%
CCS group Other gastrointestinal disorders	34775	.71%
CCS group Other perinatal conditions	34214	.70%
CCS group Crushing injury or internal injury	33283	.68%
CCS group Other upper respiratory disease	32306	.66%
CCS group Cancer of bronchus, lung	31956	.65%
CCS group Gastrointestinal haemorrhage	31834	.65%
CCS group Abdominal pain	31453	.64%
CCS group Cancer of breast	31219	.64%
CCS group Intracranial injury	29403	.60%
CCS group Intestinal obstruction without hernia	29162	.59%
CCS group Fluid and electrolyte disorders	28882	.59%
CCS group Other nutritional, endocrine, and metabolic disorders	28152	.57%
CCS group Intestinal infection	27751	.57%
CCS group Pancreatic disorders (not diabetes)	27448	.56%
CCS group Calculus of urinary tract	27337	.56%
CCS group Acute bronchitis	27261	.56%
CCS group Nonspecific chest pain	26377	.54%
CCS group Other screening for suspected conditions	24198	.49%
CCS group Asthma	24019	.49%

Abbreviations: AF atrial fibrillation; AMI acute myocardial infarction; CCS Clinical Classification Software (definition see Methods); COPD chronic obstructive pulmonary disease

Fig 1

Proportion of admissions over time for the 12 chronic conditions in different countries.

Abbreviations: AF atrial fibrillation; AMI acute myocardial infarction; COPD chronic obstructive pulmonary disease.

Overall, 43,573 (5.0%) patients died in-hospital with significant differences between the conditions (Table 3), leaving 822,929 admissions for readmission analysis. Of those, 87,452 (10.6%) had an emergency readmission within 30 days (Table 3). Readmission rate was highest in the USA (19.0%), followed by Australia (17.6%), England (16.5%) and continental Europe (14.6%). No major changes were seen over time, but there were some differences between the countries (Fig 2). Apart from anaemia and COPD, more than half of the readmissions were due to diseases other than the initial hospitalisation (Table 3). Interestingly, average LOS for the initial admission was comparable between patients that were not readmitted and were readmitted with the same condition in all 12 conditions but significantly longer in those readmitted due to another condition (Fig 3). Though the extent of this increase in LOS varied, it was seen in basically all conditions.

Table 3

Rate of in-hospital mortality, 30-day readmissions, percentage of conditions of readmissions and total events (combination of readmissions and deaths) in relation to the reason for index hospitalisation.

						% of all readmissions
Condition	Total admissions	# of deaths	In-hospital mortality	Total discharges	30-day readmissions	Same condition	Other chronic condition	Other condition	Total events	Percentage events
Pneumonia	97,496	9,959	10.2%	87,537	11,507 (13.1%)	24.1%	20.5%	55.4%	21,466	22.0%
Heart failure	84,963	5,078	6.0%	79,885	13,333 (16.7%)	42.1%	22.7%	35.2%	18,411	21.7%
COPD	63,450	2,456	3.9%	60,994	11,231 (18.4%)	51.9%	17.9%	30.2%	13,687	21.8%
Renal failure	57,209	3,101	5.4%	54,108	8,579 (15.9%)	18.4%	18.8%	62.8%	11,680	20.4%
Anemia	38,163	568	1.5%	37,595	6,818 (18.1%)	52.9%	10.2%	36.9%	7,386	19.4%
Stroke	86,189	11,105	12.9%	75,084	5,134 (6.8%)	26.8%	14.8%	58.4%	16,239	18.8%
Diabetes	41,136	580	1.4%	40,556	5,310 (13.1%)	42.6%	13.8%	43.6%	5,890	14.3%
Other cardiovascular	72,414	3,894	5.4%	68,520	6,433 (9.4%)	21.2%	17.6%	61.2%	10,327	14.3%
Hypertension	7,353	63	0.9%	7,290	620 (8.5%)	22.9%	35.7%	41.4%	683	9.3%
AMI	175,437	5,374	3.1%	170,063	10,659 (6.3%)	27.4%	27.4%	45.2%	16,033	9.1%
AF	82,723	1,326	1.6%	81,397	6,169 (7.6%)	32.2%	26.3%	41.5%	7,495	9.1%
Arthritis	59,969	69	0.1%	59,900	1,659 (2.8%)	10.9%	17.4%	71.7%	1,728	2.9%
Total	866,502	43,573	5.0%	822,929	87,452 (10.6%)	33.9%	19.9%	46.2%	131,025	15.1%

Abbreviations: AF atrial fibrillation; AMI acute myocardial infarction; COPD chronic obstructive pulmonary disease

Fig 2

Proportion of readmissions over time in different countries for each of the 12 chronic conditions.

Abbreviations: AF atrial fibrillation; AMI acute myocardial infarction; COPD chronic obstructive pulmonary disease.

Fig 3

Length of stay (LOS) for the index hospitalisation of the chronic conditions in those that were not readmitted (grey), readmitted with the same condition (orange) and with another condition (blue). Abbreviations: AF atrial fibrillation; AMI acute myocardial infarction; COPD chronic obstructive pulmonary disease.

Overall, the average and median time to readmission did not differ between readmission due to the same or another condition (Fig 4). The median time until readmission was 11 days for both the same and another condition. Average time until readmission did not vary much between the 12 conditions with only hypertension and stroke having considerably shorter time to readmission for the same condition (Fig 4).

Fig 4

Average time to readmission overall and per original condition, depending on if readmission was due to the same or another condition.

Abbreviations: AF atrial fibrillation; AMI acute myocardial infarction; COPD chronic obstructive pulmonary disease.

The specific reasons for readmission of the 12 conditions are depicted in Fig 5A and 5B. These figures show that the causes of readmissions for a reason other than the same condition are diverse for all conditions. Regarding other causes than the 12 conditions, complications related to procedures / care during the initial hospitalisation or related to devices / implants are the most common ones, followed by infections. However, there was a large diversity of other causes (Fig 5B).

Fig 5

Link between original condition of admission (left side) and cause of readmission (right side) for each chronic condition. B provides information about all other causes in A. These figures (Supporting information: Interacting Fig 1A and 1B) can be found online by clicking on the figures to provide a dynamic depiction of the different causes of readmission for each chronic condition. Abbreviations: AF atrial fibrillation; AMI acute myocardial infarction; COPD chronic obstructive pulmonary disease.

In addition to readmission for the same reason (Hellinger distance = 0), several weak clusters (light orange) were identified, such as pneumonia and COPD; HF, MI and AF; other cardiovascular, chronic renal failure and arthritis (Fig 6). COPD and anaemia had the greatest distance from each other, and anaemia was far away from all other conditions. Anaemia was also the condition with the highest proportion of readmissions with the same condition (Table 3).

Fig 6

Readmission pattern for each of the chronic conditions (Hellinger distance matrix).

Abbreviations: AF atrial fibrillation; AMI acute myocardial infarction; COPD chronic obstructive pulmonary disease.

Various factors made readmission with the same condition more or less likely (Table 4). Patients with the initial hospitalisation due to COPD, HF, anaemia, and diabetes were most likely admitted with the same condition. In the USA, readmission due to the same cause was less likely. The chance of being readmitted due to the same cause was increased in male patients, with lower age, with less co-morbidities, at the beginning of the study period and with shorter LOS during initial admission. The effect of age was particularly evident in patients aged <60 years, and less evident at older ages, but the effect of age was not the same for all chronic conditions (Fig 7).

Table 4

Odds ratios (OR) of multivariable regression for probability of readmission with the same condition as initial hospitalisation.

	OR (95%-confidence interval)
Chronic Condition: Arthritis	1.000 (Reference)
Chronic Condition: AF	4.257*** (3.610, 5.021)
Chronic Condition: AMI	3.366*** (2.863, 3.958)
Chronic Condition: Anaemia	7.328*** (6.219, 8.634)
Chronic Condition: Renal failure	2.021*** (1.713, 2.386)
Chronic Condition: COPD	9.460*** (8.057, 11.109)
Chronic Condition: Diabetes	5.468*** (4.632, 6.454)
Chronic Condition: Heart failure	7.774*** (6.624, 9.122)
Chronic Condition: Hypertension	2.485*** (1.943, 3.179)
Chronic Condition: Other cardiovascular	2.397*** (2.027, 2.834)
Chronic Condition: Pneumonia	2.934*** (2.494, 3.451)
Chronic Condition: Stroke	3.852*** (3.255, 4.559)
Country: Australia	1.000 (Reference)
Country: England	.968 (.922, 1.017)
Country: Continental Europe	.979 (.935, 1.026)
Country: USA	.728*** (.698, .760)
Female	.954*** (.926, .983)
Age	.984*** (.983, .984)
Comorbidity score	.986*** (.985, .987)
Year	.987*** (.978, .996)
LOS (sqrt days)	.870*** (.859, .881)
C-statistic	.695
Observations	87,452

*p < .1

**p < .05

***p < .01

Abbreviations: AF atrial fibrillation; AMI acute myocardial infarction; COPD chronic obstructive pulmonary disease

Fig 7

Readmission rates with the same condition by age for each of the chronic conditions.

Abbreviations: AF atrial fibrillation; AMI acute myocardial infarction; COPD chronic obstructive pulmonary disease.

Discussion

In this large multinational cohort of hospital admissions with various chronic conditions, readmissions within 30 days were common and the reason for readmission differed from the index admission in more than 50% of the cases. This is true among all participating countries and all conditions investigated, despite significant differences between countries and conditions. This is the most comprehensive study reporting readmission rates for a broad spectrum of chronic conditions in different countries. It highlights the extent and the complexity of the problem as well as the need for focus not only on the disease that caused the initial hospitalisation, but also on other frequently related conditions. This stresses the need to approach readmission prevention from a holistic consideration of all the patient’s comorbidities.

When a person-centred approach is to the whole patient, readmission rates may be reduced [24]. It is surprising that there is not more focus on this, given that the number of comorbidities is associated with increased mortality and an increased readmission rate, both early and later after discharge [25, 26]. If comorbidities primarily increase the susceptibility of deteriorating with the same condition or simply increase the risk of manifestation of other diseases has not yet been properly investigated, but comorbidities clearly increase the complexity of the condition [27]. In patients with an index hospitalisation of COPD, readmission due to other causes was accompanied with significantly higher mortality than readmission due to COPD [28]. Therefore, the patterns accompanied with the highest risks of readmission should be more adequately investigated and addressed to prevent early events. Readmission is a multifactorial phenomenon constructed as a quality measure in many countries, with contributions from underlying disease and comorbidities (as found in our data), patient factors (e.g. psychosocial resilience), healthcare worker effects, environmental and social determinants of health, organisational and healthcare system factors [15, 16, 29, 30]. In addition, the organisation of care post-discharge, such as timeliness of follow-up, coordination with primary care, and quality of medication management may significantly influence readmission rate [31, 32]. Preventable readmission is what we should be focusing on, but consensus on how to achieve this is poor [33]. Our data suggest that the focus on and treatment of the index condition only, as currently done in many institutions is not sufficient. Thus, our results explain, at least in part, why measures to prevent rehospitalisation have had only a limited impact on the readmission rate, and that successful programs must address multiple aspects of patient care [26, 30].

Disease management programs are advocated in chronic conditions such as HF or COPD after discharge to reduce the readmission rate [34-36]. Although there is no uniform way to provide managed care, the majority of such programs improve outcome as shown in several meta-analyses [37-39]. The rather broad approach to care of these programs might result in reduction of readmission not only due to the index disease. The clustering of diseases with some clinical validity found in our analysis might allow such assumption. This notion remains speculative, however, until properly tested.

Although the spectrum of causes for readmissions was broad, independently of the cause of the initial admission, there are some conditions that were more common. Thus, myocardial infarction or atrial fibrillation may lead to HF, (treatment of) HF may cause renal failure, and patients with COPD are more susceptible to acquire pneumonia. There were, however, causes that were seen in basically all conditions. Many patients had infections as cause of readmission, in line with previous reports [28]. It suggests that such patients are vulnerable early after discharge to acquire infectious diseases or infections may be acquired during hospitalisation (e.g. urinary tract infections due to urine catheters; hospital acquired pneumonia). Additionally, the susceptibility of these patients may be higher than during more stable conditions. Using sufficient hygiene measures such infections might in part be preventable if the general awareness is improved. Education of patients and their families about the increased risk of infection may also be important to aid rapid diagnosis and treatment.

Another important readmission group is related to complications of medical procedures or devices. Although it is not well investigated to what extent such complications are preventable in the setting of chronic conditions, it may be worthwhile to specifically investigate such complications and to test preventive measures to reduce the number. Often, such measures are only taken in the context of surgery and/or invasive procedures, but not as a general routine in patients with chronic conditions.

There was a small, but significant trend for less readmissions with the same condition as the index hospitalisation over time. The aging population and the fact that comorbidities are increasingly frequent at higher age may be accountable for this finding. Age and comorbidities are included in the regression equation as well, which may highlight this, but may have diminished the overtime effect. Therefore, we may expect that the complexity of patients admitted with chronic conditions is increasing in the future as the population ages further. It may also be explained as less complex procedures, and less severe conditions, are treated more in the outpatient setting. Therefore, the increasing complexity of hospitalised patients needs to be considered for future planning of hospital care of chronic diseases.

Limitations

There are several limitations to our study. Importantly, we used administrative data, capturing only what was coded. Coding is not uniform and may vary depending on diagnosis related group, degree of in-hospital testing, national reimbursement incentives, physician and coder training and institutional variation [40, 41]. Some of these factors may explain why patients in the USA had a higher comorbidity score although they were younger and had a shorter LOS. The higher comorbidities in US patients could also reflect issues of access and late presentation [42]. Direct comparison between countries, and also between institutions, regarding the absolute burden of disease is not possible. Also, risk adjustment considering all factors influencing readmissions is limited. However, the purpose of this study was to investigate the associations between index hospitalisation and readmission. As this only relates to the primary diagnosis, coding has a much lower impact. Moreover, the aim was to investigate the global picture of the problem, which is not significantly influenced by the limitation of coding. We did not compare single institutions with each other due to reasons of confidentiality given the limited number of participating hospitals, particularly in continental Europe. Administrative models may have limited discriminatory abilities, which raises concerns about the ability to standardise risk across hospitals to fairly compare hospital performance [16]. Our data suggest that previously insufficiently recognised factors related to the complexity of multimorbid patients need to be included to improve risk prediction and adjustment.

We included conditions that present with acute events (i.e. AMI, pneumonia, stroke) but are usually caused by a chronic underlying disease. Including these conditions might have influenced our results. However, readmission pattern of them were in line with the other more chronic conditions, supporting the assumption that the readmissions were mainly influenced by the underlying chronic condition rather than the acute event.

In addition, participating hospitals are academic centres and focus may differ between centres. Therefore, there may be an imbalance in the patient population and differences in the spectrum of patients in centres that did not participate. Also, readmission to a different hospital could not be recorded. Therefore, the true readmission rate is likely even higher than reported here although the disease specific readmission rates may be similar to those reported in the past (for example 7 vs 5% for heart failure) [43]. It might be speculated that readmission due to other causes than the initial admission is more likely to be treated in another hospital than the index admission and the absolute values must be interpreted with caution. Nevertheless, the consequence of these two factors would be an even higher readmission rate and a higher proportion of causes different than the initial admission, even more strongly supporting the conclusion of our study.

Conclusions

Readmissions within 30 days after discharge are very common in patients with chronic conditions. These readmissions are, in more than 50% of patients, related to a different cause than the initial hospitalisation. Therefore, this study has significant implications, for clinicians, the hospital administrators and for policy makers. It highlights the importance of a global approach to the treatment and management of patients with chronic conditions. Focus on the chronic condition and the circumstances that led to the hospitalisation is not sufficient, and additional measures based on a more holistic approach to the individual patient should be taken to significantly reduce readmissions.

List of participating centres.

(DOCX)

Click here for additional data file.

ICD 9 and ICD 10 codes for chronic conditions.

(DOCX)

Click here for additional data file.

A and B: Link between original condition of admission (left side) and cause of readmission (right side) for each chronic condition. B provides information about all other causes in A. Interactive online figures to provide a dynamic depiction of the different causes of readmission for each chronic condition.

(HTML)

Click here for additional data file.

24 Feb 2020

PONE-D-19-35626

Reasons for readmission after hospital discharge in patients with chronic diseases  - information from an international dataset

PLOS ONE

Dear Dr. Brunner-La Rocca,

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

We would appreciate receiving your revised manuscript by Apr 09 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Gianluigi Savarese

Academic Editor

PLOS ONE

Journal Requirements:

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

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at http://www.plosone.org/attachments/PLOSOne_formatting_sample_main_body.pdf and http://www.plosone.org/attachments/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. In ethics statement in the manuscript and in the online submission form, please provide additional information about the database used in your retrospective study. Specifically, please ensure that you have discussed whether all data were fully anonymized before you accessed them and/or whether the IRB or ethics committee waived the requirement for informed consent. If patients provided informed written consent to have their data used in research, please include this information.

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

"Maria Bogdanovskaya: I have read the journal's policy and the authors of this manuscript have the following competing interests: former employee of Dr Foster Telstra Health.

Lorna Barclay: I have read the journal's policy and the authors of this manuscript have the following competing interests: employee of Dr Foster Telstra Health."

We note that one or more of the authors are employed by a commercial company: Dr Foster Telstra Health.

Please provide an amended Funding Statement declaring this commercial affiliation, as well as a statement regarding the Role of Funders in your study. If the funding organization did not play a role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript and only provided financial support in the form of authors' salaries and/or research materials, please review your statements relating to the author contributions, and ensure you have specifically and accurately indicated the role(s) that these authors had in your study. You can update author roles in the Author Contributions section of the online submission form.

Please also include the following statement within your amended Funding Statement.

“The funder provided support in the form of salaries for authors [insert relevant initials], but did not have any additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. The specific roles of these authors are articulated in the ‘author contributions’ section.”

If your commercial affiliation did play a role in your study, please state and explain this role within your updated Funding Statement.

2. Please also provide an updated Competing Interests Statement declaring this commercial affiliation along with any other relevant declarations relating to employment, consultancy, patents, products in development, or marketed products, etc.

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

Please include both an updated Funding Statement and Competing Interests Statement in your cover letter. We will change the online submission form on your behalf.

Please know it is PLOS ONE policy for corresponding authors to declare, on behalf of all authors, all potential competing interests for the purposes of transparency. PLOS defines a competing interest as anything that interferes with, or could reasonably be perceived as interfering with, the full and objective presentation, peer review, editorial decision-making, or publication of research or non-research articles submitted to one of the journals. Competing interests can be financial or non-financial, professional, or personal. Competing interests can arise in relationship to an organization or another person. Please follow this link to our website for more details on competing interests: http://journals.plos.org/plosone/s/competing-interests

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

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

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

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

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

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

6. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

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

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

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

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

Reviewer #1: Yes

Reviewer #2: Partly

Reviewer #3: Yes

**********

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

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: I Don't Know

**********

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

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

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: No

**********

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

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

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

5. Review Comments to the Author

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

Reviewer #1: The authors study the reasons for readmission after hospital discharge in patients with chronic diseases in an international dataset.

In a very large sample size of 4,901,584 admissions, the authors investigated reasons for readmission in patients with an index hospitalization for multiple chronic conditions and two acute diseases usually related to an underlying chronic condition.

The hypothesis of the authors was that reasons for readmission are often different from the index hospitalisation, irrespective of the underlying condition, and that this complexity increases with age and comorbidities. In addition, the authors determined if the results changed over time.

The study is relevant, interesting and well written. I think this paper has the ability for a better understanding of the treatment and management of patients with chronic conditions and the described high rate of about 17% of readmissions. The conclusions could be strengthened. However, the present study is significantly limited by the administrative nature of the data, the lack of robust risk adjustment and the imbalances in the patient population.

There are some major concerns and limitations which have to be to addressed:

- The authors should clarify whether the identified ICD codes were main or secondary diagnoses for readmission of the patients. This fact could potentially bias the observed findings.

- Why did the authors choose the 30 days as time point of emergency readmission?

- The authors should stratify their findings concerning AMI as reason of readmission in the subtypes of AMI and by OPS codes for coronary interventions. These findings should also be incorporated in the rates of in-hospital mortality and might show differences over the time and by area (country).

- Table 4 shows the calculated odds-ratios of regression for probability of readmission with the same condition as initial hospitalisation. The authors should adjust the findings i.e. by area (country), length of stay and year of admission.

Reviewer #2: In the present paper, Brunner-La Rocca and Colleagues aimed to "investigate reasons for readmission in patients with an index hospitalisation for multiple chronic conditions and two acute diseases usually related to an underlying chronic condition”. The Authors collected electronic inpatient records from administrative discharge data provided by each of the 21 participating hospitals in the years 2009-2014. The final analysis was performed on around 866k admissions due to 12 different chronic conditions. 30-day readmission rates ranged from 2.8% for arthritis to 18.4% for COPD; the Authors also report a high incidence of re-admission due to other (non-chronic) conditions, most commonly, treatment related complications and infections.

The issue of readmission rate in chronic disease is of outmost social and economic importance, but it has been often regarded from a disease-specific point of view. The Authors should be commended for their attempt to provide a picture of the possible paths patients with major chronic conditions often follow, using a huge amount of administrative data from all over the world.

While of interest, the value of the findings of the paper is limited from some major methodological issues, as reported below in details.

Major points

- The Authors state that they “included two acute conditions, i.e. AMI and bacterial pneumonia”. Indeed, stroke (acute?!) was also considered in the analysis. Although they were apparently included as they underly other chronic conditions, the Reviewer is rather concerned about this choice.

- Only readmissions at the same hospital could be recorded. Although the disease specific readmission rates are similar to those reported in the past (for example 7 vs 5% for heart failure, Fudim M et al. Aetiology, timing and clinical predictors of early vs. late readmission following

index hospitalization for acute heart failure: insights from ASCEND-HF. Eur J Heart Fail. 2018;20:304-314), the whole number of readmission may have been significantly underestimated, especially as concerns those caused by different diseases (as patients may have been referred to different specialized hospitals).

- Administrative and reimbursement issues may have significantly influenced the attribution of the readmission cause, thus possibly biasing the Authors’ findings.

- How was time to first readmission related to the readmission cause?

Minor points

- How do the Authors explain their observation that readmission rated have not changed significantly along the whole study period for most of the chronic conditions they have considered?

- Discussion should be shortened and focused on the main findings of the study.

Reviewer #3: Thank you for the opportunity to review this manuscript. It explored causes of admission and readmission in a unique population of multiple academic centers across the world. In a type of consortium 20+ centers have pooled administrative data for purposes of quality improvement research. IN this analysis the researchers found that in most cases the cause of readmission is different from the cause of admission the first time around. These findings have been previously described in a variable type of cohorts including administrative databases and clinical trial cohorts (see my last comment). Nevertheless this data remains very important and only underlines the need for a more “holistic” patient care (as the authors called it).

Comments:

- This study encompasses multiple countries. US policy on readmission reduction was cited but what about other countries? Do they have comparable initiatives?

- I am not sure I follow the difference between ICD and CCS. Aka Table 1 and 2.

- It is important to understand where the ICD code for the primary diagnosis was derived from. Did the ICD code get obtained from the billing information, discharge notes etc? Only the code I the first position was used?

- Found the amount of information provide in the figures to be somewhat overwhelming.

- Abbreviations in all tables and figures need to be presented in legends.

- Information in the discussion section is repetitive. Reduce repetition and reduce in length.

- Given that HF is a major contributor to the admission and readmission consider to discuss some of the already published literature on this topic. PMID: 27133201 and PMID: 29082629.

**********

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

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

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

Reviewer #1: No

Reviewer #2: No

Reviewer #3: No

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

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

12 Apr 2020

Response to editorial comments:

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at http://www.plosone.org/attachments/PLOSOne_formatting_sample_main_body.pdf and http://www.plosone.org/attachments/PLOSOne_formatting_sample_title_authors_affiliations.pdf

Ok.

2. In ethics statement in the manuscript and in the online submission form, please provide additional information about the database used in your retrospective study. Specifically, please ensure that you have discussed whether all data were fully anonymized before you accessed them and/or whether the IRB or ethics committee waived the requirement for informed consent. If patients provided informed written consent to have their data used in research, please include this information.

Indeed, the IRB waived the requirements of informed consent. This information is included in the revised version of the manuscript.

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

"Maria Bogdanovskaya: I have read the journal's policy and the authors of this manuscript have the following competing interests: former employee of Dr Foster Telstra Health.

Lorna Barclay: I have read the journal's policy and the authors of this manuscript have the following competing interests: employee of Dr Foster Telstra Health."

We note that one or more of the authors are employed by a commercial company: Dr Foster Telstra Health.

1. Please provide an amended Funding Statement declaring this commercial affiliation, as well as a statement regarding the Role of Funders in your study. If the funding organization did not play a role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript and only provided financial support in the form of authors' salaries and/or research materials, please review your statements relating to the author contributions, and ensure you have specifically and accurately indicated the role(s) that these authors had in your study. You can update author roles in the Author Contributions section of the online submission form.

Please also include the following statement within your amended Funding Statement.

“The funder provided support in the form of salaries for authors [insert relevant initials], but did not have any additional role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. The specific roles of these authors are articulated in the ‘author contributions’ section.”

If your commercial affiliation did play a role in your study, please state and explain this role within your updated Funding Statement.

2. Please also provide an updated Competing Interests Statement declaring this commercial affiliation along with any other relevant declarations relating to employment, consultancy, patents, products in development, or marketed products, etc.

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

Please include both an updated Funding Statement and Competing Interests Statement in your cover letter. We will change the online submission form on your behalf.

Please know it is PLOS ONE policy for corresponding authors to declare, on behalf of all authors, all potential competing interests for the purposes of transparency. PLOS defines a competing interest as anything that interferes with, or could reasonably be perceived as interfering with, the full and objective presentation, peer review, editorial decision-making, or publication of research or non-research articles submitted to one of the journals. Competing interests can be financial or non-financial, professional, or personal. Competing interests can arise in relationship to an organization or another person. Please follow this link to our website for more details on competing interests: http://journals.plos.org/plosone/s/competing-interests

“Competing interests related to the involvement of two authors need to be declared. MB and LB had an active role in the analysis of the data and participated in the design of the analysis of this manuscript. The company (Dr Foster Telstra Health) collected the data, but did not influence the study design and the data analysis. The company also did not have any influence on the decision to publish or the preparation of the manuscript. The involvement of two employees of the company does not alter our adherence to PLOS ONE policies on sharing data and materials. However, the data cannot be made available publicly as indicated below.” This statement is included in the competing interest section.

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

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

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

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

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

Data cannot be made publicly available as the terms and conditions of this collaboration restricted the availability of original data as follows:

1) Only certain categories of people are permitted to access and use the “Analytical Tool” (that is, the software processing input data and generating output data for providing benchmarking to each hospital for quality improvements). These are defined as:

“Permitted Users” means the Participant’s directors, employees and independent contractors who are authorised by the Participant to access and use the Analytical Tool for the purposes provided in this Agreement.

2) Access to and use of the Analytical Tool itself is granted only for the “Term” and for specific purposes, being “accessing and viewing the Output Data and to use the Output Data for the purpose of participating in and reviewing the results of the Project”. The restriction is related to the contract between the participating hospitals and Dr. Foster Telstra Health. There it was specified who has access to the data and that they may not be shared with other people as described above. Other researchers can request access to the data as long as they fulfil the requirements of the contract. Contracting the responsible person in each individual hospital is possible, but the authors cannot guarantee that each hospital will provide the data. This is the responsibility of each hospital. It would be all administrative data of the according years in the participating hospitals. The data are owned by the participating hospitals. A list of participating hospitals, [and their contact information] is included in the supporting information.

The following hospitals were participating in this study:

Hospital name City Country Contact Website

Alfred Health Melbourne Australia +61 3 90762000 www.alfredhealth.org.au

Austin Health Melbourne Australia +61 3 94965000 www.austin.org.au

Melbourne Health Melbourne Australia +61 3 93427000 www.thermh.org.au

Monash Health Melbourne Australia +61 3 95946666 monashhealth.org

Universitair ziekenhuis Leuven Leuven Belgium +32 16 332211 www.uzleuven.be/en

Aalborg UH Aalborg Denmark +45 97 666000 aalborguh.rn.dk/service/english

Guy's and St Thomas' London England +44 20 75895111 www.imperial.ac.uk

Imperial College Healthcare London England +44 20 33113311 www.imperial.nhs.uk

Royal United Hospitals Bath Bath England +44 1225 428331 www.ruh.nhs.uk

University College London Hospitals London England +44 20 34567890 www.uclh.nhs.uk

University Hospitals Coventry & Warwickshire Coventry England +44 24 76964000 www.uhcw.nhs.uk

Academisch Ziekenhuis Maastricht Maastricht The Netherlands +31 43 3876543 www.mumc.nl

Humanitas Research Hospital Milan Italy +39 02 82246250 www.humanitas.it

Barnes-Jewish Hospital St. Louis USA +1 314 7473000 www.barnesjewish.org

Hackensack University Medical Center Hackensack USA +1 844 4649355 www.hackensackumc.org

Hospital of the University of Pennsylvania Philadelphia USA +1 215 6624000 www.pennmedicine.org

Huntsville Hospital Huntsville USA +1 256 2651000 www.huntsvillehospital.org

Keck Hospital of USC Los Angeles USA +1 800 8722273 www.keckmedicine.org

UC San Diego Medical Center San Diego USA +1 858 6577000 health.ucsd.edu

University of Texas Southwestern Medical Center Dallas USA +1 214 6483111 www.utsouthwestern.edu

Yale–New Haven Hospital New Haven USA +1 203 6884242 www.ynhh.org

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

The according sentence (“Analysing only 2011-2013 did not affect results (data not shown).”) has been removed and is no longer part of the revised manuscript.

6. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

We have added the description of the link to the interactive figures 5A and 5B to the legend of figure 5 (previous figure 4), which now reads as follows:

“Figure 5 Link between original condition of admission (left side) and cause of readmission (right side) for each chronic condition. Figure B provides information about all other causes in figure A. These figures (Supporting information: Interacting Figures 4A and 4B) can be found online by clicking on the figures to provide a dynamic depiction of the different causes of readmission for each chronic condition”

In addition, we provided the following information at the end of the manuscript:

“Supplementary table 1: List of participating centres

Supplementary table 2: ICD 9 and ICD 10 codes for chronic conditions

Supplementary figures 5A and 5B: Link between original condition of admission (left side) and cause of readmission (right side) for each chronic condition. Figure B provides information about all other causes in figure A. Interactive online figures to provide a dynamic depiction of the different causes of readmission for each chronic condition”

Comments to comments by the reviewers

Reviewer #1:

The study is relevant, interesting and well written. I think this paper has the ability for a better understanding of the treatment and management of patients with chronic conditions and the described high rate of about 17% of readmissions. The conclusions could be strengthened. However, the present study is significantly limited by the administrative nature of the data, the lack of robust risk adjustment and the imbalances in the patient population.

We would like to thank the reviewer for carefully reading our manuscript and to provide positive feedback on it.

We agree with the reviewer that the study has important limitations that we also addressed in part in the limitation section of the original submission. We now added the fact that risk adjustment was limited and that there may be an imbalance in the patient population. This is inevitable to administrative data. The advantage of administrative data is that there is no preselection of patients, which is inevitable to most registries. For the purpose of the study, we are convinced that administrative data are useful.

There are some major concerns and limitations which have to be to addressed:

- The authors should clarify whether the identified ICD codes were main or secondary diagnoses for readmission of the patients. This fact could potentially bias the observed findings.

We agree with the reviewer. This is why we only used main diagnosis for readmission of patients to keep the potential bias as small as possible. We are aware that the coding may be subjective and not homogenous across all centres. This is addressed in the limitation section of the manuscript.

- Why did the authors choose the 30 days as time point of emergency readmission?

To some extent, this is obviously arbitrary. However, the reason for this is mainly twofold: a) reimbursement policy in the USA is related to the 30-day readmission rate. The result of this fact is that 30-day readmission is often used in publications, which makes comparison more accurate; b) we wanted to avoid interference with new diseases that develop over time to really be able to focus on readmissions related to the initial diagnosis. We did not include this in the revised manuscript as we do not think that this is a crucial part for the understanding. However, if the editor feels that we should add a short paragraph to highlight these thoughts, we are certainly willing to do so.

- The authors should stratify their findings concerning AMI as reason of readmission in the subtypes of AMI and by OPS codes for coronary interventions. These findings should also be incorporated in the rates of in-hospital mortality and might show differences over the time and by area (country).

Unfortunately, this information is not available with sufficient accuracy. Due to countries having different coding versions, the coding is not sufficient across all countries and sites for this analysis. This is an important reason why we did not split diagnoses / readmissions in further details. Moreover, it would not change the frequency of readmissions due to causes other than the index event. Thus, the main message of the study is not influenced by this. Still, we agree with the reviewer that this would be interesting in a study that specifically addresses the issues of readmissions in AMI patients. However, this is beyond the scope of this study.

- Table 4 shows the calculated odds-ratios of regression for probability of readmission with the same condition as initial hospitalisation. The authors should adjust the findings i.e. by area (country), length of stay and year of admission.

We agree with the reviewer that this is interesting, which is why we did the multivariable regression analysis as depicted in table 4. Thus, the odds ratios provided are adjusted by all the variables included. Obviously, this was not clear which is why we added that the table contains the results of multivariable regression.

Reviewer #2:

The issue of readmission rate in chronic disease is of outmost social and economic importance, but it has been often regarded from a disease-specific point of view. The Authors should be commended for their attempt to provide a picture of the possible paths patients with major chronic conditions often follow, using a huge amount of administrative data from all over the world.

We would like to thank the reviewer for his/her positive comment on our study.

While of interest, the value of the findings of the paper is limited from some major methodological issues, as reported below in details.

Major points

- The Authors state that they “included two acute conditions, i.e. AMI and bacterial pneumonia”. Indeed, stroke (acute?!) was also considered in the analysis. Although they were apparently included as they underly other chronic conditions, the Reviewer is rather concerned about this choice.

We agree with the reviewer that stroke is also an acute event, based on an underlying chronic disease. We changed this accordingly in the revised manuscript.

Moreover, we understand the concern by the reviewer. In fact, we had some discussion about this point when designing the analysis. We decided to do so as we aimed to provide a broad picture of readmission of important different chronic diseases and that the underlying disease is most important for them. Also the other included conditions usually have acute events that lead to hospitalisation (e.g. acute decompensation of heart failure or CODP, occurrence of atrial fibrillation). In fact, the readmission rate of these three acute conditions were in line with the other conditions, supporting our choice. Therefore, we did not change the diseases included in the analysis. Nevertheless, we included the following paragraph to the limitation section:

“We included conditions that present with acute events (i.e. AMI, pneumonia, stroke) but are usually caused by a chronic underlying disease. Including these conditions might have influenced our results. However, readmission pattern of them were in line with the other more chronic conditions, supporting the assumption that the readmissions were mainly influenced by the underlying chronic condition rather than the acute event.”

- Only readmissions at the same hospital could be recorded. Although the disease specific readmission rates are similar to those reported in the past (for example 7 vs 5% for heart failure, Fudim M et al. Aetiology, timing and clinical predictors of early vs. late readmission following index hospitalization for acute heart failure: insights from ASCEND-HF. Eur J Heart Fail. 2018;20:304-314), the whole number of readmission may have been significantly underestimated, especially as concerns those caused by different diseases (as patients may have been referred to different specialized hospitals).

We fully agree with the reviewer which is why we included this discussion point in the original limitation section. In addition, we added the comparison with heart failure as an example as suggested by the reviewer.

- Administrative and reimbursement issues may have significantly influenced the attribution of the readmission cause, thus possibly biasing the Authors’ findings.

We fully agree with the reviewer. This is why we mention this point as our first limitation to our study. As also highlighted in one of the points raised by reviewer #1, there are pros and cons for the use of administrative data for this kind of analyses. For the purpose of our study, we are convinced that the inclusion of all cases of a specific hospital is an important pro, allowing the conclusions made.

- How was time to first readmission related to the readmission cause?

This is an interesting point. In fact, the time to readmission was exactly the same if comparing similar or different cause of readmission overall. There were small differences between the different original conditions. Due to the large number of patients, they obviously are all statistically significantly different, but of clinical relevance, only two are meaningful. We added this information to the manuscript in a separate paragraph of the Results section and added a figure (new figure 4).

Minor points

- How do the Authors explain their observation that readmission rated have not changed significantly along the whole study period for most of the chronic conditions they have considered?

We tend to disagree with the conclusion by the reviewer that there were no changes over time. In fact, there was a significant reduction in admission due to the same condition over time (1.3% per year) as indicated in table 4. As discussed, we hypothesise that this is related to the increase in age / complexity of chronic diseases over time.

- Discussion should be shortened and focused on the main findings of the study.

We agree in part with the reviewer and have shorten the discussion. We are also aware of the fact that our data are not sufficiently in depth to address all questions related to the topic. In order to stimulate the discussion and additional research in the field – both to better understand and to prevent readmissions – we deliberately discussed some issues where our results are more hypothesis generating than providing clear evidence.

Reviewer #3:

Thank you for the opportunity to review this manuscript. It explored causes of admission and readmission in a unique population of multiple academic centers across the world. In a type of consortium 20+ centers have pooled administrative data for purposes of quality improvement research. IN this analysis the researchers found that in most cases the cause of readmission is different from the cause of admission the first time around. These findings have been previously described in a variable type of cohorts including administrative databases and clinical trial cohorts (see my last comment). Nevertheless this data remains very important and only underlines the need for a more “holistic” patient care (as the authors called it).

We would like to thank the reviewer for the positive comment on our study.

Comments:

- This study encompasses multiple countries. US policy on readmission reduction was cited but what about other countries? Do they have comparable initiatives?

Although each healthcare system aims to reduce readmission rates, the US policy in this regard is rather unique. This is why we focused on discussing this in our manuscript. Given the fact that reduction in the length of the discussion is suggested by this and another reviewer, we have not extended our discussion in this point.

- I am not sure I follow the difference between ICD and CCS. Aka Table 1 and 2.

The ICD code defines diseases in more detail, where as the CCS summarises diseases in groups. It is referenced in the paper (see Methods section, ref 17)

- It is important to understand where the ICD code for the primary diagnosis was derived from. Did the ICD code get obtained from the billing information, discharge notes etc? Only the code I the first position was used?

Administrative dataset contain data that are used for national registers and are obtained from discharge notes. As mentioned in the manuscript, only the main diagnosis was used.

- Found the amount of information provide in the figures to be somewhat overwhelming.

We understand this notion by the reviewer. However, we are convinced that this is a strength of the manuscript as information is available if readers are interested in the details of our findings. This is why we did not reduce it.

- Abbreviations in all tables and figures need to be presented in legends.

We agree with the reviewer and changed this accordingly.

- Information in the discussion section is repetitive. Reduce repetition and reduce in length.

We agree in part with the reviewer and have shorten the discussion, particularly to avoid repetition unless required to understand the context of the according paragraph. We are also aware of the fact that our data are not sufficiently in depth to address all questions related to the topic. In order to stimulate the discussion and additional research in the field – both to better understand and to prevent readmissions – we deliberately discussed some issues where our results are more hypothesis generating than providing clear evidence.

- Given that HF is a major contributor to the admission and readmission consider to discuss some of the already published literature on this topic. PMID: 27133201 and PMID: 29082629.

We agree with the reviewer that HF is an important chronic condition. As also suggested by one of the other reviewers, we added one of the two suggested references to the manuscript, mentioning HF as example for comparison with our results.

Submitted filename: Rebuttal PLOS ONE.docx

Click here for additional data file.

6 May 2020

Reasons for readmission after hospital discharge in patients with chronic diseases  - information from an international dataset

PONE-D-19-35626R1

Dear Dr. Brunner-La Rocca,

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

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

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

With kind regards,

Gianluigi Savarese

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: All comments have been addressed

Reviewer #3: All comments have been addressed

**********

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

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

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

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

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

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

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

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

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

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

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

6. Review Comments to the Author

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

Reviewer #1: Dear Editor,

the authors were able to answer all queries extensively and sufficiently.

In my opinion the present study explored causes of admission and readmission in a unique population.

From the reviewer´s view, the manuscript can be accepted for publication in PLOS ONE.

With best regards and thanks in advance,

Moritz Becher

Reviewer #2: The Authors have addressed, as far as possible, the issues raised in the previous revision of the manuscript.

Reviewer #3: Thank you for addressing my comments. I have no further suggestions at this time.

..................

**********

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

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

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

Reviewer #1: No

Reviewer #2: No

Reviewer #3: Yes: Marat Fudim

20 May 2020

PONE-D-19-35626R1

Reasons for readmission after hospital discharge in patients with chronic diseases  - information from an international dataset

Dear Dr. Brunner-La Rocca:

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

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

For any other questions or concerns, please email plosone@plos.org.

Thank you for submitting your work to PLOS ONE.

With kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Gianluigi Savarese

Academic Editor

PLOS ONE