Prospective controlled study that recruited patients from outpatient's respiratory clinics after being discharged from the hospital due to AECOPD in a tertiary teaching hospital between 2012 and 2019. Patients were identified after two or more hospital admissions per year due to AECOPD and were offered to join a fe-COPD ICM if they fulfilled the following inclusion criteria: (1) patients with diagnosis of COPD according to GOLD criteria,1 able to attend scheduled visits to hospital, (2) able to contact the ICM team by phone and, (3) able to follow the treatment instructions. Patients were excluded if they (1) had non-respiratory comorbidity that substantially affected the prognosis of the disease, (2) lived out of the area of influence of the hospital, (3) lived in a nursing home or jail, (4) suffered from deteriorative cognitive function or psychiatric illness that affected their mental capacity, or (5) had socio-familial conditions that affected the access to the hospital or to attend the scheduled visits. The control group were similarly COPD patients diagnosed according to GOLD criteria with history of ≥2 AECOPD required hospitalization and fulfilled inclusion criteria with no exclusion criteria for ICM but they were unwilling to be included in the ICM, or refused to contact the program team through the phone and preferred to come directly to ER, or preferred to continue with their primary physician of care. The control group underwent standard care in the same hospital with scheduled outpatient visits in Primary and Specialized Care and were included in a database as controls and analyzed retrospectively. Exacerbations were defined, and all patients received treatment, according to GOLD recommendations.1 The study was approved by the research board of the participating hospital and all the participants signed a written informed consent.

The fe-COPD patients enrolled in the ICM were assigned to a nurse-led program under medical supervision that included personalized-structured visits with social, nutritional, rehabilitation, educational (including smoking cessation when required) and functional assessment. Also, patients were provided with a fast-track access to medical care in case of new respiratory symptoms by a direct telephone call with the ICM team. The evaluation and plans with every patient were shared with their Primary Care nurse through a web-based chronic disease management package included in the local health care electronic record system.

At first visit, patients were subjected to: (1) complete medical history with assessment of comorbidities and evaluation of health status with COPD assessment test (CAT), the modified Medical Research Council (mMRC) dyspnea scale, smoking history, and history of previous exacerbations; (2) forced spirometry, lung volumes and DLCO (if there was not a previous one within the 6 months prior to admission to the program), ADO index (age, dyspnea and airway obstruction), 6-minute walking test, arterial blood gases and BODE index was calculated; (3) sputum sample for microbiological evaluation; (4) revision of inhalation techniques and compliance by pharmacy refills in the electronic prescription records; (5) education about healthy habits, alert signs of exacerbations and aiming at behavior changes including smoking cessation; (6) nutritional status and need of dietary assessment; (7) physical activities and need of rehabilitation; and (8) assessment of psychosocial support and identification care-givers with coaching with patient and family support. All patients were followed-up on regular basis at scheduled visits every 3 months (or earlier if required) with evaluation of symptoms, vital signs, mMRC dyspnea scale, CAT score, O2 saturation and re-checking of inhalation techniques by specialized nurse at each single visit. Identification and treatment of new comorbidities, compliance of treatments, physical activity and psychosocial needs were also assessed in every scheduled visit. Moreover, patients included in the program were able to contact the ICM team from Monday-to-Friday through telephone calls for consultation about their symptoms outside the scheduled visits. Based on their new presenting symptoms, the team decided whether the patient needed to come to hospital for medical evaluation and accordingly further action was considered (online supplement fig. 1). Out-of-hours, patients requiring urgent medical attention had to go to the ER at their Primary Care center or hospital.

New hospital admissions, domiciliary hospitalization, ER visits, courses of antibiotic and systemic corticosteroids as well as mortality were recorded for every patient during follow-up for further analysis. All variables were analyzed as cumulative variables throughout the whole follow-up duration not as annual frequency. Average days of hospitalization were recorded for each patient and analyzed. Patients with AECOPD who were managed by the ICM team as outpatient were considered as avoided ER visits. When patients fulfilled criteria for hospitalization, providing that they were not on acute respiratory failure, and were conscious with good home support, they were managed as outpatient instead with close monitoring by the program team. This was considered as avoided admissions. Baseline and follow-up values of mMRC and CAT were collected prospectively in the ICM group.

The data were expressed as median and interquartile range (IQR) or number and percentage (%) as appropriate unless otherwise stated. Mann–Whitney, unpaired t-test and chi-square tests were used in the comparison between both groups as appropriate. A percentage of reduction of new hospital admissions and ER visits was calculated to study the effect of the ICM as the relation between measured new hospital admission or ER visits/expected hospitalizations or emergency visits (i.e. actual new hospital admission or emergency visits+avoided admissions or ER visits).

C-statistic was used to identify the best cut-off point for CAT score as predictor of mortality among fe-COPD ICM patients. Logistic univariate and multivariate analysis were used to evaluate the risk factors for new hospital admissions and mortality among fe-COPD ICM patients. Odd ratio (OR) and confidence interval 95% (CI95%) were calculated. Cumulative mortality within 7 years of enrollment in the ICM was estimated using Kaplan–Meier survival analysis. A two-tailed p value<0.05 was considered statistically significant. SPSS package (Version 22.0. Armonk, NY: IBM Corp) was used for all analyses.

280 fe-COPD patients were followed-up for a median of 37 months (IQR=14–65 months) and maximally 7 years. 148 fe-COPD patients were assigned to fe-COPD ICM, and 132 fe-COPD were unwilling to join the program and considered as a control group receiving standard medical care. Seven fe-COPD from the ICM group (4.7%) abandoned the program due to the occurrence of exclusion criteria during follow-up; accordingly, 141 patients completed the program and were analyzed (Fig. 1).

Fig. 1.

Flow chart of the studied population.

(0.09MB).

Table 1 shows the baseline clinical and functional characteristics of the studied population. All included fe-COPD patients were predominantly GOLD D in both groups with similar proportion of patients with airway colonization by Pseudomonas aeruginosa between both groups. Further, the ADO index was similar between both groups (p=0.550, Table 1). Noteworthy, patients in the fe-COPD ICM group were slightly younger (72 (IQR=66.0–77.0) vs 76 (IQR=69.5–83.0) years), with more severe airway obstruction and lower BMI than the control group (Table 1). Seventy-five patients (56.8%) of control group were on domiciliary oxygen therapy versus 104 patients (73.8%) of ICM group (p=0.003); while 23.5% versus 21.3% respectively used nocturnal non-invasive ventilation (p>0.05, Table 1).

The effectiveness of integrated care models has been previously demonstrated in uncontrolled studies. Huertas et al.15 found that a day hospital-based care model for severe COPD patients was associated with a reduction of hospital admission by 71%. Similarly, Jain et al.18 found that there was significant decrease of ER visits and hospitalizations in asthma and COPD patients with frequent exacerbations in a similar ICM. On the other hand, other forms of integrated care using home telemonitoring rather than face-to-face evaluation in frequent COPD exacerbators also reported significant decrease of ER visits, hospital admissions and the number of exacerbations/year.13 Our results are in accordance with these studies; however, our study had the strength of providing a control group of fe-COPD. Further, the previous mentioned studies had an average follow-up of 12 months13,15,18 and maximum of 24 months,15 whereas in the current study, the median follow-up duration was 31 months with maximum of 7 years.

A key objective of COPD management is preventing exacerbations, that have a significant burden on the quality of life especially among severe COPD patients.12,19 According to our results, we believe that enrolment of fe-COPD in an integrated comprehensive program is associated with significant decrease of AECOPD that required ER visits, hospitalization, and, in case patients needed to be admitted, a reduction of hospitalization days.

Several factors may explain the effectiveness of ICM in the current study: (1) the flexibility of the program that enabled the patient to access to medical care through a direct telephone call and so the early detection of exacerbations and subsequent management; (2) rapid arrangement of follow-up visits after phone calls for evaluation of patients after exacerbations; (3) the health care is provided by the same team who are familiar with the patients’ characteristics; (4) the program is conducted in a tertiary hospital with all facilities for rapid tracking and evaluation of the patients; (5) the care model is coordinated with Primary Care and other hospital departments such as Hospital-at-home, Rehabilitation, Nutrition or social assistant. All of these factors also increased the adherence of the patients to the ICM visits and the given instructions which positively enhanced the impact of the program.

However, despite the impact of this model of care on hospital utilization, we did not find differences in mortality when comparing to standard care. This could be explained on the basis that fe-COPD patients enrolled in the ICM were more severe in terms of lower FEV1 and higher basal mMRC dyspnea scale compared to control group (although there was a similar proportion of GOLD D and C in both groups as well as ADO index). We found that ADO index was a significant risk factor of mortality in the entire fe-COPD population that was not superior than mMRC dyspnea scale (online supplement Table 3). Other studies have shown that basal mMRC dyspnea scale is a good predictor of mortality and AECOPD.20 mMRC was also a good predictor of mortality in our cohort that was not influenced by the ICM. However, dyspnea is a complex symptom that can be explained by different pathophysiological mechanisms, especially in a more severe and frequent-exacerbator population like ours. When we evaluated a multidimensional score such as the BODE index, an important predictor of survival among stable COPD irrespective of their ABCD grouping,21 we found that it was a significant risk factor of mortality. However, in our population of fe-COPD receiving ICM, BODE index was not superior to CAT score for the prediction of mortality. CAT is a simple questionnaire that has been validated as a multi-dimension assessment tool for COPD patients and can be used for the evaluation of treatment response.22 Further, there was a significant decrease of CAT score after 1 year and 3 years of follow-up among fe-COPD patients enrolled in the program denoting better control of the disease and improving health status. Not surprising, we found that last follow-up CAT score>17 was the strongest predictor for new admissions and mortality among fe-COPD. CAT score>17 has been used as cutoff value for the identification of poor health status and moderate-severe exacerbations leading to death.20,23 Rassouli et al.24 also reported a significant positive association between CAT score changes and risk of AECOPD. In our understanding, this finding may have important prognostic implications since the ICM have a direct impact on CAT.

The high mortality rate observed in the studied COPD population could be explained by the severity of the underlying COPD disease especially FEV125, but also by patients’ demographics. Morevover, COPD is associated to a persistent underlying systemic inflammatory process that could be a mediator of extra-pulmonary complications and a risk factor for mortality.26,27 Moreover, still there is a percentage of COPD who die during their sleep or in whom death was unexpected.28

Interestingly, we found that female gender was associated to mortality in our fe-COPD cohort. Ringbaek et al.29 found that women with COPD had double risk for mortality compared to men. Further, Stolz et al.30 found that females with COPD were at greater risk for moderate-to-severe AECOPD than males. Moreover, Pseudomonas aeruginosa infection was also a significant risk factor for mortality but other bacterial infections were not. Pseudomonas aeruginosa infection was shown to be a predictor of 3-year mortality after hospitalization for AECOPD31 and in outpatient stable COPD patients.32 However, age, gender or Pseudomonas aeruginosa infection were not independent risk factors in the multivariate analysis among fe-COPD included in the ICM program, which we think that can be explained by the interactions among all these factors on the severity of the disease.

The current study has some limitations. Firstly, there could be a selection bias since the control group was not randomly assigned. However, the population enrolled in the ICM was more severe than the control group which reinforces the conclusions. Moreover we thought that it was unethical to choose a randomized control group when we have a specialized program of care for this group of patients. Secondly, the data of hospitalizations, AECOPD and treatments of the control group were obtained from the recorded files of the hospital; however, we assumed that there was no significant bias in the analysis as all the patients were recruited and followed up in the same hospital. Thirdly, we did not analyze the impact of the ICM in terms of costs. The saving on hospital admissions and ER visits should be counterbalance with the cost of having a structure with a full-time nurse and a respiratory specialist on demand. Since the facilities and personnel were already in place and were reallocated to exert this function, we anticipate that the cost/benefit was high for our institution, although detailed economic analysis is needed. Previous studies considered telemonitoring of COPD had significant reduction of total costs for COPD management.33,34 Fourth, the current program was applied in a single center and all the patients were living in the same geographical area; however, this fact improved the adherence to the program, so enhanced better control of the patients. Lastly, we did not analyze social factors and physical activity as risk factors among our studied population. Physical activity level is considered as an important factor associated with mortality and exacerbations in COPD population,35 however, our ICM program is integrated and personalized offering rehabilitation and social support to our patients as part of their usual care.