JMIR Publications

ABSTRACT

Background:

Hospital patient registries provide substantial longitudinal datasets describing the clinical and medical health statuses of inpatients and their pharmacological prescriptions. Despite the multiple advantages of routinely collecting multidimensional longitudinal data, those datasets are rarely suitable for advanced statistical analysis and they require customisation and synthesis.

Objective:

To describe the methods used to transform and synthesise a raw, multidimensional, hospital patient registry dataset into an exploitable database for investigating risk profiles and predictive and survival health outcomes among polymorbid, polymedicated, older inpatients in relation to their medicines prescriptions at hospital discharge.

Methods:

A raw, multidimensional dataset from a public hospital was extracted from the hospital registry in a CSV file and imported into the R statistical package for cleaning, customisation and synthesis. Patients fulfilling the criteria for inclusion were home-dwelling, polymedicated, older adults with multiple chronic conditions (HDOAs) aged 65 years old or more who became hospitalised. The patient dataset covered 140 variables from 20,422 HDOA-hospitalisations from 2015 to 2018. Each variable, according to type, was explored and computed to describe distributions, missing values and associations. Different clustering methods, expert opinion, recoding and missing-value techniques were used to customise and synthesise these multidimensional datasets.

Results:

Sociodemographic data showed no missing values. Average age, hospital length of stay and frequency of hospitalisation were computed. Discharge details were recoded and summarised. Clinical data were cleaned up and best practices for managing missing values were applied. Seven clusters of medical diagnoses, surgical interventions, somatic, cognitive and medicines data were extracted using empirical and statistical best practices, with each presenting the health status of the subjects included in them as accurately as possible. Medical, comorbidity and drug data were recoded and summarised.

Conclusions:

A cleaner, better-structured dataset was obtained, combining empirical and best-practice statistical approaches. The overall strategy delivered an exploitable, population-based database suitable for an advanced analysis of the descriptive, predictive and survival statistics relating to HDOAs admitted as inpatients. More research is needed to develop best practices for customising and synthesising large, multidimensional, population-based registries.