A systematic review of risk stratification tools internationally used in primary care settings

Abstract

Background and Aims In our current healthcare situation, burden on healthcare services is increasing, with higher costs and increased utilization. Structured population health management has been developed as an approach to balance quality with increasing costs. This approach identifies sub‐populations with comparable health risks, to tailor interventions for those that will benefit the most. Worldwide, the use of routine healthcare data extracted from electronic health registries for risk stratification approaches is increasing. Different risk stratification tools are used on different levels of the healthcare continuum. In this systematic literature review, we aimed to explore which tools are used in primary healthcare settings and assess their performance. Methods We performed a systematic literature review of studies applying risk stratification tools with health outcomes in primary care populations. Studies in Organisation for Economic Co‐operation and Development countries published in English‐language journals were included. Search engines were utilized with keywords, for example, “primary care,” “risk stratification,” and “model.” Risk stratification tools were compared based on different measures: area under the curve (AUC) and C‐statistics for dichotomous outcomes and R 2 for continuous outcomes. Results The search provided 4718 articles. Specific election criteria such as primary care populations, generic health utilization outcomes, and routinely collected data sources identified 61 articles, reporting on 31 different models. The three most frequently applied models were the Adjusted Clinical Groups (ACG, n = 23), the Charlson Comorbidity Index (CCI, n = 19), and the Hierarchical Condition Categories (HCC, n = 7). Most AUC and C‐statistic values were above 0.7, with ACG showing slightly improved scores compared with the CCI and HCC (typically between 0.6 and 0.7). Conclusion Based on statistical performance, the validity of the ACG was the highest, followed by the CCI and the HCC. The ACG also appeared to be the most flexible, with the use of different international coding systems and measuring a wider variety of health outcomes.

Care delivery team composition effect on hospitalization risk in African Americans with congestive heart failure

“…Current computational algorithms for CHF include the: Adjusted Clinical Groups, Charlson Comorbidity, Elixhauser Comorbidity Indices, Hierarchical Condition Categories, and Framingham Risk Score [ 34 – 36 ]. Broadly, current tools rely on aggregated patient demographic data (e.g., age, sex, race, marital status), comorbidities (e.g., number and type of chronic conditions), patient panels (e.g., glucose, lipid, systolic and diastolic levels), and behavioral (e.g., diet, physical activity, smoking) to compute a single risk score that assigns patients into risk tiers [ 16 , 17 , 34 – 36 ]. However, in solely relying on such indicators [ignoring care delivery influences], current tools have not optimized their utility in comprehensively approaching RS [ 13 , 28 , 37 ].…”

Expert panel identified health-related social needs and methodological considerations for a polysocial risk score

“…The aggregated, summative nature of a risk score combined with challenges in measurement timing, could conflict with physicians' and staffs' need to know specific individual-level HRSNs and pressing needs in order to adjust care or to provide assistance directly or through referrals. Instead, population-level awareness would be conceptually closer to risk stratification 44,45 where organizations would be able to identify larger patient groups for additional screening or for follow-up activities Regardless, bias in inferences about patients and misuse of the score are paramount concerns to the expert panel as summary risk scores for numerous conditions and risks have been demonstrated to be biased 48,49 . Selective reporting and collecting of HRSNs creates distorted picture of risk 50 .…”

Using deep-learning to obtain calibrated individual disease and ADL damage transition probabilities between successive ELSA waves

“…We find HRs between 3 and 10 (the largest possible for deciles) – indicating substantial risk stratification is possible with our approach. Our well-calibrated model provides accuracy in risk prediction results which would provide reliability for decision-making [34, 35]. Since we had calibrated probabilities, we also considered the distributions of transition probabilities.…”