Selecting indicators for the measurement of low-value care using German claims data: A three-round modified Delphi panel

By reducing healthcare services that offer little benefit or potential harm to patients (low-value care), resources can be redirected towards more adequate treatments, improving healthcare efficiency and patient outcomes. This study aimed to systematically incorporate clinical expertise across medical disciplines through a Delphi process to establish indicators for measuring low-value care, ensuring their acceptance by medical societies, the broader medical community, and patients. We developed two versions (one with higher sensitivity and one with higher specificity) for almost each of the 42 indicators identified as potentially measurable in a previous systematic review. We conducted a three-round modified Delphi panel based on the RAND/UCLA appropriateness methodology, with 62 experts from 52 Scientific Medical Societies and professional organisations, and patient representatives. In round one, each indicator was rated for its ability to indicate low-value healthcare and its measurability in German claims data. This was followed by an online discussion in round two. The indicators were then modified based on expert feedback and re-assessed in round three. As a result, 24 indicators were deemed appropriate for measuring low-value care, covering areas such as pharmaceuticals, diagnostic tests, screening, and treatment. For example, one indicator identified patients with cancer who received chemotherapy in the last month of life. These indicators will help identify healthcare services that may require policy-level interventions to improve the quality of care. However, most low-value care indicators can only be measured in German claims data if documentation requirements for relevant information are expanded.