Development and validation of predictive models for unplanned hospitalization in the Basque Country: analyzing the variability of non-deterministic algorithms.

BACKGROUND: The progressive ageing in developed countries entails an increase in multimorbidity. Population-wide predictive models for adverse health outcomes are crucial to address these growing healthcare needs. The main objective of this study is to develop and validate a population-based prognostic model to predict the probability of unplanned hospitalization in the Basque Country, through comparing the performance of a logistic regression model and three families of machine learning models. METHODS: Using age, sex, diagnoses and drug prescriptions previously transformed by the Johns Hopkins Adjusted Clinical Groups (ACG) System, we predict the probability of unplanned hospitalization in the Basque Country (2.2 million inhabitants) using several techniques. When dealing with non-deterministic algorithms, comparing a single model per technique is not enough to choose the best approach. Thus, we conduct 40 experiments per family of models - Random Forest, Gradient Boosting Decision Trees and Multilayer Perceptrons - and compare them to Logistic Regression. Models' performance are compared both population-wide and for the 20,000 patients with the highest predicted probabilities, as a hypothetical high-risk group to intervene on. RESULTS: The best-performing technique is Multilayer Perceptron, followed by Gradient Boosting Decision Trees, Logistic Regression and Random Forest. Multilayer Perceptrons also have the lowest variability, around an order of magnitude less than Random Forests. Median area under the ROC curve, average precision and positive predictive value range from 0.789 to 0.802, 0.237 to 0.257 and 0.485 to 0.511, respectively. For Brier Score the median values are 0.048 for all techniques. There is some overlap between the algorithms. For instance, Gradient Boosting Decision Trees perform better than Logistic Regression more than 75% of the time, but not always. CONCLUSIONS: All models have good global performance. The only family that is consistently superior to Logistic Regression is Multilayer Perceptron, showing a very reliable performance with the lowest variability.

Gender and socioeconomic inequalities in the implementation of the Basque programme for multimorbid patients.

BACKGROUND: The aim of our study was to increase awareness of the relevance of the implemented programmes to inequity of access and inequality of health by analyzing the impact of a patient-centred strategy for multimorbid patients. METHODS: This retrospective study compared the 2014 multimorbid patient group (intervention group) with its 2012 analogue (control group), before the Department of Health of the Basque Country launched the strategy for managing disease chronicity. Inequalities in healthcare access were represented by differences in the inclusion of patients in the programme and in contacts with primary care (PC) services by gender and socioeconomic status (measured by deprivation index by census track). Likewise, differences in hospital care represented inequalities in health outcomes. Generalized linear models were used to analyze relationships among variables. A propensity score by a genetic matching approach was used to minimize possible selection bias. RESULTS: At baseline, women had less probability of being eligible for the programme. No clear patterns were seen in resource consumption in PC. The probability of hospitalization was higher for men and increased according to socioeconomic status. The implementation of the programme yielded more contacts with PC services in all groups and a reduction in hospitalizations, especially among men and the most socioeconomically deprived patients. CONCLUSION: The patient-centred, integrated-care intervention launched by the Department of Health of the Basque Country might have reduced some gender and socioeconomic inequalities in health outcomes, as it avoided more hospitalizations in subgroups that presented with more episodes of decompensation in the reference year.

Social inequalities in a population based colorectal cancer screening programme in the Basque Country.

BACKGROUND: While it is known that a variety of factors (biological, behavioural and interventional) play a major role in the health of individuals and populations, the importance of the role of social determinants is less clear. The effect of social inequality on population-based screening for colorectal cancer (CRC) could limit the value of such programmes. The present study aims to determine whether such inequalities exist. METHODS: Data was obtained from the population-based screening programme administered in the Autonomous Community of the Basque Country, Spain, with a target population aged 50 to 69, first invited to participate between 2009 and 2011. The magnitude of inequality was analysed using the odds ratio (taking the least disadvantaged socioeconomic quintile as the reference population), the population attributable risk and the relative index of inequality, based on the regression, which is the ratio of the rates in the most and least disadvantaged socioeconomic groups. RESULTS: The target population comprised 242,394 people, with the test kit successfully sent to 95.1 % (230,510). The overall response rate was 64.3 % (67.1 in women and 61.4 % men). Among women, the highest participation was in the third quintile (71.5 %) and the lowest in the first - the least disadvantaged (65.7 %). The lowest and highest rates of people with identified lesions were in the second and fourth quintiles (14.7/1000 and 17.0/1000 respectively). Among men, the response rate was lowest in the fifth - most disadvantaged - quintile (60.2 %). The highest rate of identified lesions was in the fifth quintile; 38 % higher than the first (55.7/1000 compared to 41.0/1000). CONCLUSIONS: Sex and socioeconomic group influence the rate of participation in the CRC programme and the rate of lesions found in the participants. Any public health programme is morally and ethically obliged to strive for equity and effectiveness. Improving participation of men and socially disadvantaged groups should be taken in account.

Is integration of healthy lifestyle promotion into primary care feasible? Discussion and consensus sessions between clinicians and researchers.

BACKGROUND: The adoption of a healthy lifestyle, including physical activity, a healthy diet, moderate alcohol consumption and abstinence from smoking, is associated with a major decrease in the incidence of chronic diseases and mortality. Primary health-care (PHC) services therefore attempt, with rather limited success, to promote such lifestyles in their patients. The objective of the present study is to ascertain the perceptions of clinicians and researchers within the Basque Health System of the factors that hinder or facilitate the integration of healthy lifestyle promotion in routine PHC setting. METHODS: Formative research based on five consensus meetings held by an expert panel of 12 PHC professionals with clinical and research experience in health promotion, supplied with selected bibliographic material. These meetings were recorded, summarized and the provisional findings were returned to participants in order to improve their validity. RESULTS: The Health Belief Model, the Theory of Planned Action, the Social Learning Theory, "stages of change" models and integrative models were considered the most useful by the expert panel. Effective intervention strategies, such as the "5 A's" strategy (assess, advise, agree, assist and arrange) are also available. However, none of these can be directly implemented or continuously maintained under current PHC conditions. These strategies should therefore be redesigned by adjusting the intervention objectives and contents to the operation of primary care centres and, in turn, altering the organisation of the centres where they are to be implemented. CONCLUSION: It is recommended to address optimisation of health promotion in PHC from a research perspective in which PHC professionals, researchers and managers of these services cooperate in designing and evaluating innovative programs. Future strategies should adopt a socio-ecological approach in which the health system plays an essential role but which nevertheless complements other individual, cultural and social factors that condition health. These initiatives require an adequate theoretical and methodological framework for designing and evaluating complex interventions.

Cross-national health care database utilization between Spain and France: results from the EPICHRONIC study assessing the prevalence of type 2 diabetes mellitus.

AIM: The EPICHRONIC (EPIdemiology of CHRONIC diseases) project investigated the possibility of developing common procedures for French and Spanish electronic health care databases to enable large-scale pharmacoepidemiological studies on chronic diseases. A feasibility study assessed the prevalence of type 2 diabetes mellitus (T2DM) in Navarre and the Basque Country (Spain) and the Midi-Pyrénées region (France). PATIENTS AND METHODS: We described and compared database structures and the availability of hospital, outpatient, and drug-dispensing data from 5.9 million inhabitants. Due to differences in database structures and recorded data, we could not develop a common procedure to estimate T2DM prevalence, but identified an algorithm specific to each database. Patients were identified using primary care diagnosis codes previously validated in Spanish databases and a combination of primary care diagnosis codes, hospital diagnosis codes, and data on exposure to oral antidiabetic drugs from the French database. RESULTS: Spanish and French databases (the latter termed Système National d'Information Inter-Régimes de l'Assurance Maladie [SNIIRAM]) included demographic, primary care diagnoses, hospital diagnoses, and outpatient drug-dispensing data. Diagnoses were encoded using the International Classification of Primary Care (version 2) and the International Classification of Diseases, version 9 and version 10 (ICD-9 and ICD-10) in the Spanish databases, whereas the SNIIRAM contained ICD-10 codes. All data were anonymized before transferring to researchers. T2DM prevalence in the population over 20 years was estimated to be 6.6-7.0% in the Spanish regions and 6.3% in the Midi-Pyrénées region with ~2% higher estimates for males in the three regions. CONCLUSION: Tailored procedures can be designed to estimate the prevalence of T2DM in population-based studies from Spanish and French electronic health care records.

Subclinical depressive symptoms and continued cannabis use: predictors of negative outcomes in first episode psychosis.

BACKGROUND: Although depressive symptoms in first episode psychosis have been associated with cannabis abuse, their influence on the long-term functional course of FEP patients who abuse cannabis is unknown. The aims of the study were to examine the influence of subclinical depressive symptoms on the long-term outcome in first episode-psychosis patients who were cannabis users and to assess the influence of these subclinical depressive symptoms on the ability to quit cannabis use. METHODS: 64 FEP patients who were cannabis users at baseline were followed-up for 5 years. Two groups were defined: (a) patients with subclinical depressive symptoms at least once during follow-up (DPG), and (b) patients without subclinical depressive symptoms during follow-up (NDPG). Psychotic symptoms were measured using the Positive and Negative Syndrome Scale (PANSS), depressive symptoms using the Hamilton Depression Rating Scale (HDRS)-17, and psychosocial functioning was assessed using the Global Assessment of Functioning (GAF). A linear mixed-effects model was used to analyze the combined influence of cannabis use and subclinical depressive symptomatology on the clinical outcome. RESULTS: Subclinical depressive symptoms were associated with continued abuse of cannabis during follow-up (ß= 4.45; 95% confidence interval [CI]: 1.78 to 11.17; P = .001) and with worse functioning (ß = -5.50; 95% CI: -9.02 to -0.33; P = .009). CONCLUSIONS: Subclinical depressive symptoms and continued cannabis abuse during follow-up could be predictors of negative outcomes in FEP patients.

[Evaluation of healthcare quality for acute cerebrovascular disease in the context of the service level agreement]. / Evaluación de la calidad de la atención a la enfermedad cerebrovascular aguda en el marco del contrato programa.

BACKGROUND: The service level agreement establishes the quality requirements for those services contracted by the Basque Government Health Department from Osakidetza-Servicio Vasco de Salud. Acute cerebrovascular disease (ACVD) is one of the care processes with quality specifications. OBJECTIVE: To outline the procedure adopted to evaluate ACVD care and the results obtained in four hospitals in Vizcaya between 2003 and 2007. MATERIAL AND METHODS: In 2003, a work group consisting of clinicians, hospital and quality assurance managers, and experts from the Regional Health Board chose a series of indicators which would be measured by an external auditor. This group reviews the results annually and sets objectives for the following financial year. RESULTS: The improvement in almost all the indicators has been significant, and those with high ratings from the beginning have maintained their level. The percentage of patients who had a CAT scan in the first six hours after arriving at hospital increased from 57% to 85%, and the administration of anticoagulants within 12 hours increased from 70% of cases to 90%. The fibrinolysis rate was 3.8% in two hospitals. The percentage of patients who began rehabilitation whilst admitted was less than 3%. CONCLUSIONS: After operating for five years, the procedure adopted, which involves collaboration between clinical, management and planning staff, has been shown to be viable and effective in improving the quality of ACVD care.