JMIR Publications

ABSTRACT

Background:

The inter-contact interval (ICI), defined as the number of days between consecutive practice contacts, measures individual patient contact behavior but requires further validation as a useful tool for primary care research.

Objective:

This study applies the ICI to different electronic datasets from primary care practices to characterize individual patients’ contact patterns over time. The ICI and electronic “big data” are then used to develop and refine measures that may signal increased contact frequency and help identify acute care needs.

Methods:

This retrospective longitudinal observational study used electronic medical records (EMRs) from German primary care practices for a secondary data analysis, comprising (1) all patients from 155 primary care practices over 14 years, (2) all patients from one practice over 24 years, and (3) 100 patients receiving oral anticoagulation from seven practices over 7 years. ICI distributions were summarized using means, standard deviations, medians, 25th and 75th percentiles, interquartile ranges, and boxplots for a random sample. Ordinal time series analysis was used to identify increased practice contact frequency and ICI minima that best reflected a potential new need for acute care and patterns of contact behavior.

Results:

Among 333,610 patients with ≥2 practice contacts, 7,044,653 ICIs were identified across the three datasets. Patients had practice contacts, on average, every 38, 53, and 15 days, respectively. Nearly half of the patients had at least one contact at a 1-day interval every 3 months. ICI frequency distributions consistently showed peaks at 7 days and their multiples and at 90–93 days. Substantial variation in ICI distributions was observed within individual patients, between patients, between practices, and across datasets. Time series analyses and further refinements suggested defining a short ICI minimum as a contact with an ICI shorter than those of its direct neighbors and <8 days, potentially indicating acute care needs. Low ICI values, indicating frequent contacts, were not randomly distributed; instead, they frequently formed clusters (“islands”) of acute care within longer periods of chronic care. Patients could be divided into three groups: patients with a nearly constant consultation rhythm and no urgent practice visit; patients with 1 or more periods of intensified contact, suggesting acute care needs; and patients with frequent short-interval practice contacts.

Conclusions:

While a high number of contacts is a simple indicator of potentially intensified acute care needs, refined measures, such as ICI minima provide a more stringent indication of care intensity and the severity of acute care needs. Ordinal time series analysis and graphical ICI analysis can identify “islands” of frequent attendance and distinguish acute care episodes from the surrounding chronic care trajectory, demonstrating the usefulness of the ICI as a tool for longitudinal analyses of real-world primary care data. Clinical Trial: n.a.