COPD Readmissions Prediction- ReCOPD Project

Overview

To develop predictive models for short-term readmission (60 days) and medium term (one year) for patients with an exacerbation of COPD (eCOPD) which are admitted to any of the participant hospitals. To identify factors contributing to readmission from the time of discharge from the index admission to the studied readmission time. Methodology. Design: Prospective cohort study with a nested case control study, with follow up to one year. Patients: 1500 patients admitted over a year and a half in any of the 9 participant hospitals by an eCOPD. Methods: During admission, background information on patient situation prior to admission, data of the ED during admission, and discharge planning, will be collected from medical history. The investigators will also survey patients (generic and specific health-related quality life -HRQoL- dependence, social support, presence of fragility). All patients will be followed to determining whether any readmission until 60 and up to one year of discharge from the index admission occurs. Patients readmitted at the first 60 days after discharge will be surveyed on a number of possible causes of readmission (continuity of care, medication, complications, individual and social factors). A one to one control group will be established to compare different factors studied in the case group (readmission). Statistical analysis: predictive models will be developed from the derivation sample, and validated in a validation sample. Also, the factors evaluated in the case-control study will be compared using appropriate test for paired samples.

DESIGN: A prospective observational cohort study with a nested case control study, with one year follow-up from index hospital admission. SETTING: Multicenter and multidisciplinary study taking part the following hospitals: Costa del Sol Hospital (Andalusia), Nuestra Señora de la Candelaria University Hospital (Canary Islands), Gregorio Marañón General University Hospital (Madrid), and Basque Country hospitals (Araba, Donostia, Basurto, Cruces University Hospitals and Santa Marina and Galdakao-Usansolo Hospitals). SUBJECTS: In this study episodes will be considered from the index episode of the admission of the COPD exacerbated recruited patient. Inclusion criteria: episodes of patients with exacerbation of chronic obstructive pulmonary disease (COPD). Those patients can be admitted with 2 presentations: A) known case of COPD: episodes of patients with a previous diagnosis of COPD (FEV1/FVC\<70%) who have an exacerbation of COPD. The exacerbation of COPD is defined as an acute worsening of symptoms in relation to the baseline situation and beyond the patient's daily variability. In addition, the exacerbation makes necessary to make changes in the patient's usual medication. The most frequently reported symptoms are dyspnea, cough, increased the sputum's production, and changes in the sputum's color. B) New case of COPD: episodes in which COPD is suspected because of the medical history and anamnesis (smokers or ex-smorkers of more than 15 pack/year, with basal dyspnea, cough, and expectoration for more than three months per year) and it is compatible with an exacerbation. The diagnosis will be confirmed at respiratory outpatient clinic within two months after discharge, by performing a spirometry when the patient is stable. If COPD is not confirmed in this consultation, the patient would be excluded from the study. Patients who sign informed consent. Exclusion criteria: patients admitted with obstructive or restrictive ventilatory failure due to another disease (asthma, consequences of tuberculosis, pachypleuritis, restrictive diseases). All COPD patients finally admitted for any other cause, other than exacerbation or complicated COPD. Patients with severe physical or psychopathological disease, cognitive deterioration, any neurologic disease, that prevented them from properly completing the questionnaires. Patients who do not understand the Spanish language to complete properly the questionnaires. Patients that do not want to participate or do not sign the inform consent. For re-admissions: those who are programmed admissions or those for another cause not in relation with CMS algorithms. SAMPLE SIZE: Calculation of sample size: predictive modeling studies establish that it is necessary to have at least 10 events of the dependent variable of interest (readmissions) for each independent variable included in the multivariate model. In the multivariate logistic regression models, for the assessment of not less than 5 but not more than 10 variables, the investigators estimate that it will be necessary at least 100 events of the dependent variable (Readmissions) in the sample from which the investigators will derive the prediction rule to ensure that the model converges properly. With a 20% of readmissions, and requiring two samples (derivation and validation) the investigators need at least 1000 valid admissions. On the other hand, in the nested case-control study, to detect differences in the continuity of care at follow-up, from 70 to 55%, or in conciliation with medication from 30 to 15%, separately for case (readmissions) and control groups, less than 200 valid patients would be needed for the analysis in each group (for and α=0.05 and a power (1-β) of 0.80), which suppose 1000 index admissions. To do this, each center will have to collect approximately 100-250 patients during the recruitment period. This is assured, while assuming a 20% of exclusions, including those cases that are programmed admissions, and 15% of losses on the remaining sample. Sampling: consecutive, all patients who re-enter and renter during the study period. METHODS: Variables to be included in the study: Variables to be collected from the Emergency Department: 1) level of awareness measured by Glasgow scale, baseline and admission dyspnea (MRC dyspnea and Borg fatigue scales), respiratory and cardiac frequency, blood pressure, fever, thoracic pain, gasometric parameters (arterial oxygen (PO2) or oxygen saturation (spoO2), pH, arterial carbon dioxide level (PCO2)); Variables to be collected during admission: 2) Background and sociodemographic (age, sex); health habits (smoking, exercise level), hearing or visual deficits, or falls; social support; 3) baseline COPD severity (measured by FEV1, 5), emergency visits in the last month and year, and previous admissions during the last 12 months; 4) Associated diseases that affect the exacerbation of COPD (cardiac, diabetes), and all those that the Charlson Comorbidity index; 5)Weight, height (BMI); 6)electrocardiography data, radiological findings, biochemical data and hematocrit, gasometry, FIO2 on arrival and discharge; 7) Other: basal treatment in the Emergency Department, and in the admission (short and/or long-term beta inhalers, anticholinergics, inhaled and/or oral orticoid, oxygen therapy), antibiotics, and other medications necessary for comorbidities; ICU admission, invasive mechanical ventilation; 8) presence of complications, or death; Variables to be collected at admission and until discharge: ECG, radiography at discharge, physiological data (temperature, respiratory and cardiac frequency, blood preassure), biochemical data and hematocrit, gasometry and/or pulsimetry. Length of hospital stay. Non-invasive ventilation or oxygen therapy at home. Health care habits (tobacco, diet, exercise), specialist or family physician assessment, and, if available, a review of inhalers using technique. Self-administered questionnaires to be completed by the patient during admission * Assessment of the quality of life. EuroQol-5d generic questionnaire; specific questionnaire: COPD Assessment test (CAT). * Symptoms´ assessment: MRC Dyspnea Scales; Physical Activity Questionnaire made ad hoc. * Anxiety/depression assessment: HAD questionnaire. * Social dependency: -Duke-UNC Social Support Scale * Fragility assessment: Tilburg Frailty Indicator Part B (Components of frailty) * Activities of daily living: Barthel's test. Variables to be collected from 2 months and until one year of follow-up: through the participants hospitals' computer systems (electronic medical history) and/or medical record on paper, and of the national death index it will be collected basic information of the new readmissions (cause, date, duration and complications) or patient's death. DATA COLLECTED. A) Patients recruitment: trained personnel, with help/supervision of the hospitals' respiratory specialist, will consecutively and prospectively recruit patients at each center that meet the inclusion criteria. For each patient, the data collector will be introduced to each patient to inform about the study aims', and if the patient agree to participate, and sign the inform consent, HRQoL questionnaires will be given. The questionnaires should be completed by the patient itself or with the help of a family member or caregiver. The data collector will establish when to get back the questionnaires. If the patients have been discharged, the questionnaires will be mailed. B) Clinical data collection: Information will be collected from the Emergency Department and later, from the medical record during admission and up to discharge. C) Identification of readmissions in the first two months: The same data collector described above will examine the readmission from the discharge to 60 days from it, to detect patients who are re-admitted. This will be identified as readmission, and simultaneously, a patient that will be act as control and who has been admitted at the same week as the patient with the current readmission will be searched, trying to match both patients by age (by categories of five years) and baseline FEV1 (greater or less than 50). All patients who have been readmitted within two months, and their established controls, will be immediately mailed the follow-up questionnaire (the corresponding questionnaire according if is case-readmission- or control situation). They will be phoned in advance to advise them of the mailing of the questionnaires. The questionnaires will be hand-delivered in the case of the patient hospital admitted at that time. If within 7/10 days the questionnaire has not been received, a reminder letter will be sent, and if in another 7/10 days the questionnaire has not been yet received, a letter with the questionnaire will it be sent again. D) One year follow-up: Of all the patients recruited in the index admission, whether they were readmissions, controls or not, clinical data will be completed up to the year. E) Blood sample: In three of the hospital participant a sample of blood has been collected in each index admission and stored in each Biobank. In these cases, the patient must sign a specific additional informed consent established by the Biobank. With each Biobank the method will be established for the recruitment and delivery of samples.