Population segmentation based on healthcare needs: a systematic review

Impact of family doctor system on diabetic patients with distinct service utilisation patterns: a difference-in-differences analysis based on group-based trajectory modelling

INTRODUCTION

This study examines the impact of China's family doctor system (FDS) on healthcare utilisation and costs among diabetic patients with distinct long-term service utilisation patterns.

METHODS

Conducted in City A, eastern China, this retrospective cohort study used data from the Health Information System and Health Insurance Claim Databases, covering diabetic patients from 1 January 2014 to 31 December 2019.Patients were categorised into service utilisation trajectories based on quarterly outpatient visits to community health centres (CHCs) and secondary/tertiary hospitals from 2014 to 2017 using group-based trajectory models. Propensity score matching within each trajectory group matched FDS-enrolled patients (intervention) with non-enrolled patients (control). Difference-in-differences analysis compared outcomes between groups, with a SUEST test for cross-model comparison. Outcomes included outpatient visits indicator, costs indicator and out-of-pocket (OOP) expenses.

RESULTS

Among 17 232 diabetic patients (55.21% female, mean age 62.85 years), 13 094 were enrolled in the FDS (intervention group) and 4138 were not (control group). Patients were classified into four trajectory groups based on service utilisation from 2014 to 2017: (1) low overall outpatient utilisation, (2) high CHC visits, (3) high secondary/tertiary hospital visits and (4) high overall outpatient utilisation. After enrolled in FDS From 2018 to 2019, the group with high secondary/tertiary hospital visits saw a 6.265 increase in CHC visits (225.4% cost increase) and a 3.345 decrease in hospital visits (55.5% cost reduction). The high overall utilisation group experienced a 4.642 increase in CHC visits (109.5% cost increase) and a 1.493 decrease in hospital visits. OOP expenses were significantly reduced across all groups.

CONCLUSION

The FDS in China significantly increases primary care utilisation and cost, while reducing hospital visits and costs among diabetic patients, particularly among patients with historically high hospital usage. Policymakers should focus on enhancing the FDS to further encourage primary care usage and improve chronic disease management.

Improving uptake of population health management through scalable analysis of linked electronic health data

Population Health Management - often abbreviated to PHM - is a relatively new approach for healthcare planning, requiring the application of analytical techniques to linked patient level data. Despite expectations for greater uptake of PHM, there is a deficit of available solutions to help health services embed it into routine use. This paper concerns the development, application and use of an interactive tool which can be linked to a healthcare system's data warehouse and employed to readily perform key PHM tasks such as population segmentation, risk stratification, and deriving various performance metrics and descriptive summaries. Developed through open-source code in a large healthcare system in South West England, and used by others around the country, this paper demonstrates the importance of a scalable, purpose-built solution for improving the uptake of PHM in health services.

Comparing natural language processing representations of coded disease sequences for prediction in electronic health records

OBJECTIVE

Natural language processing (NLP) algorithms are increasingly being applied to obtain unsupervised representations of electronic health record (EHR) data, but their comparative performance at predicting clinical endpoints remains unclear. Our objective was to compare the performance of unsupervised representations of sequences of disease codes generated by bag-of-words versus sequence-based NLP algorithms at predicting clinically relevant outcomes.

MATERIALS AND METHODS

This cohort study used primary care EHRs from 6 286 233 people with Multiple Long-Term Conditions in England. For each patient, an unsupervised vector representation of their time-ordered sequences of diseases was generated using 2 input strategies (212 disease categories versus 9462 diagnostic codes) and different NLP algorithms (Latent Dirichlet Allocation, doc2vec, and 2 transformer models designed for EHRs). We also developed a transformer architecture, named EHR-BERT, incorporating sociodemographic information. We compared the performance of each of these representations (without fine-tuning) as inputs into a logistic classifier to predict 1-year mortality, healthcare use, and new disease diagnosis.

RESULTS

Patient representations generated by sequence-based algorithms performed consistently better than bag-of-words methods in predicting clinical endpoints, with the highest performance for EHR-BERT across all tasks, although the absolute improvement was small. Representations generated using disease categories perform similarly to those using diagnostic codes as inputs, suggesting models can equally manage smaller or larger vocabularies for prediction of these outcomes.

DISCUSSION AND CONCLUSION

Patient representations produced by sequence-based NLP algorithms from sequences of disease codes demonstrate improved predictive content for patient outcomes compared with representations generated by co-occurrence-based algorithms. This suggests transformer models may be useful for generating multi-purpose representations, even without fine-tuning.

How predictive of future healthcare utilisation and mortality is data-driven population segmentation based on healthcare utilisation and chronic condition comorbidity?

BACKGROUND

In recent years data-driven population segmentation using cluster analyses of mainly health care utilisation data has been used as a proxy of future health care need. Chronic conditions patterns tended to be examined after segmentation but may be useful as a segmentation variable which, in combination with utilisation could indicate severity. These could further be of practical use to target specific clinical groups including for prevention. This study aimed to assess the ability of data-driven segmentation based on health care utilisation and comorbidities to predict future outcomes: Emergency admission, A&E attendance, GP practice contacts, and mortality.

METHODS

We analysed record-linked data for 412,997 patients registered with GP practices in 2018-19 in Cwm Taf Morgannwg University Health Board (CTM UHB) area within the Secure Anonymised Information Linkage (SAIL) Databank. We created 10 segments using k-means clustering based on utilisation (GP practice contacts, prescriptions, emergency and elective admissions, A&E and outpatients) and chronic condition counts for 2018 using different variable compositions to denote need. We assessed the characteristics of the segments. We employed a train/test scheme (80% training set) to compare logistic regression model predictions with observed outcomes on follow-up in 2019. We assessed the area under the ROC curve (AUC) for models with demographic variables, with and without the segments, as well as between segmentation implementations (with/without comorbidity and primary care data).

RESULTS

Adding the segments to the model with demographic covariates improved the prediction for all outcomes. For emergency admissions this increased discrimination from AUC 0.65 (CI 0.64-0.65) to 0.73 (CI 0.73-0.74). Models with the segments only performed nearly as well as the full models. Excluding comorbidity showed reduced predictive ability for mortality (similar otherwise) but most pronounced reduction when excluding all primary care variables.

CONCLUSIONS

This shows that the segments have satisfactory predictive ability, even for varied outcomes and a broad range of events and conditions used in the segmentation. It suggests that the segments can be a useful tool in helping to identify specific groups of need to target with anticipatory care. Identification may be refined with selected diagnoses or more specialised tools such as risk stratification.

Application of artificial intelligence in active assisted living for aging population in real-world setting with commercial devices - A scoping review

BACKGROUND

The aging population is steadily increasing, posing new challenges and opportunities for healthcare systems worldwide. Technological advancements, particularly in commercially available Active Assisted Living devices, offer a promising alternative. These readily accessible products, ranging from smartwatches to home automation systems, are often equipped with Artificial Intelligence capabilities that can monitor health metrics, predict adverse events, and facilitate a safer living environment. However, there is no review exploring how Artificial Intelligence has been integrated into commercially available Active Assisted Living technologies, and how these devices monitor health metrics and provide healthcare solutions in a real-world environment for healthy aging. This review is essential because it fills a knowledge gap in understanding AI's integration in Active Assisted Living technologies in promoting healthy aging in real-world settings, identifying key issues that require to be addressed in future studies.

OBJECTIVE

The aim of this overview is to outline current understanding, identify potential research opportunities, and highlight research gaps from published studies regarding the use of Artificial Intelligence in commercially available Active Assisted Living technologies that assists older individuals aging at home.

METHODS

A comprehensive search was conducted in six databases-PubMed, CINAHL, IEEE Xplore, Scopus, ACM Digital Library, and Web of Science-to identify relevant studies published over the past decade from 2013 to 2024. Our methodology adhered to the PRISMA extension for scoping reviews to ensure rigor and transparency throughout the review process. After applying predefined inclusion and exclusion criteria on 825 retrieved articles, a total of 64 papers were included for analysis and synthesis.

RESULTS

Several trends emerged from our analysis of the 64 selected papers. A majority of the work (39/64, 61%) was published after the year 2020. Geographically, most of the studies originated from East Asia and North America (36/64, 56%). The primary application goal of Artificial Intelligence in the reviewed literature was focused on activity recognition (34/64, 53%), followed by daily monitoring (10/64, 16%). Methodologically, tree-based and neural network-based approaches were the most prevalent Artificial Intelligence algorithms used in studies (32/64, 50% and 31/64, 48% respectively). A notable proportion of the studies (32/64, 50%) carried out their research using specially designed smart home testbeds that simulate the conditions in real-world. Moreover, ambient technology was a common thread (49/64, 77%), with occupancy-related data (such as motion and electrical appliance usage logs) and environmental sensors (indicators like temperature and humidity) being the most frequently used.

CONCLUSION

Our results suggest that Artificial Intelligence has been increasingly deployed in the real-world Active Assisted Living context over the past decade, offering a variety of applications aimed at healthy aging and facilitating independent living for the older adults. A wide range of smart home indicators were leveraged for comprehensive data analysis, exploring and enhancing the potentials and effectiveness of solutions. However, our review has identified multiple research gaps that need further investigation. First, most research has been conducted in controlled testbed environments, leaving a lack of real-world applications that could validate the technologies' efficacy and scalability. Second, there is a noticeable absence of research leveraging cloud technology, an essential tool for large-scale deployment and standardized data collection and management. Future work should prioritize these areas to maximize the potential benefits of Artificial Intelligence in Active Assisted Living settings.

Segmentation and fragmentation of health systems and the quest for universal health coverage: conceptual clarifications from the Mexican case

Health systems are complex entities. The Mexican health system includes the private and public sectors, and subsystems that target different populations based on corporatist criteria. Lack of unity and its consequences can be better understood using two concepts, segmentation and fragmentation. These reveal mechanisms and strategies that impede progress toward universality and equity in Mexico and other low- and middle-income countries. Segmentation refers to separation of the population by position in the labour market. Fragmentation refers to institutions, and to financial aspects, health care levels, states' systems of care, and organizational models. These elements explain inequitable allocation of resources and packages of health services offered by each institution to its population. Overcoming segmentation will require a shift from employment to citizenship as the basis for eligibility for public health care. Shortcomings of fragmentation can be avoided by establishing a common package of guaranteed benefits. Mexico illustrates how these two concepts characterize a common reality in low- and middle-income countries.

Learning prevalent patterns of co-morbidities in multichronic patients using population-based healthcare data

The prevalence of longstanding chronic diseases has increased worldwide, along with the average age of the population. As a result, an increasing number of people is affected by two or more chronic conditions simultaneously, and healthcare systems are facing the challenge of treating multimorbid patients effectively. Current therapeutic strategies are suited to manage each chronic condition separately, without considering the whole clinical condition of the patient. This approach may lead to suboptimal clinical outcomes and system inefficiencies (e.g. redundant diagnostic tests and inadequate drug prescriptions). We develop a novel methodology based on the joint implementation of data reduction and clustering algorithms to identify patterns of chronic diseases that are likely to co-occur in multichronic patients. We analyse data from a large adult population of multichronic patients living in Tuscany (Italy) in 2019 which was stratified by sex and age classes. Results demonstrate that (i) cardio-metabolic, endocrine, and neuro-degenerative diseases represent a stable pattern of multimorbidity, and (ii) disease prevalence and clustering vary across ages and between women and men. Identifying the most common multichronic profiles can help tailor medical protocols to patients' needs and reduce costs. Furthermore, analysing temporal patterns of disease can refine risk predictions for evolutive chronic conditions.

The impact of dual VA-Medicare use on a data-driven clinical management tool for older Veterans with multimorbidity

BACKGROUND

Healthcare systems are increasingly turning to data-driven approaches, such as clustering techniques, to inform interventions for medically complex older adults. However, patients seeking care in multiple healthcare systems may have missing diagnoses across systems, leading to misclassification of resulting groups. We evaluated the impact of multi-system use on the accuracy and composition of multimorbidity groups among older adults in the Veterans Health Administration (VA).

METHODS

Eligible patients were VA primary care users aged ≥65 years and in the top decile of predicted 1-year hospitalization risk in 2018 (n = 558,864). Diagnoses of 26 chronic conditions were coded using a 24-month lookback period and input into latent class analysis (LCA) models. In a random 10% sample (n = 56,008), we compared the resulting model fit, class profiles, and patient assignments from models using only VA system data versus VA with Medicare data.

RESULTS

LCA identified six patient comorbidity groups using VA system data. We labeled groups based on diagnoses with higher within-group prevalence relative to the average: Substance Use Disorders (7% of patients), Mental Health (15%), Heart Disease (22%), Diabetes (16%), Tumor (14%), and High Complexity (10%). VA with Medicare data showed improved model fit and assigned more patients with high accuracy. Over 70% of patients assigned to the Substance, Mental Health, High Complexity, and Tumor groups using VA data were assigned to the same group in VA with Medicare data. However, 41.9% of the Heart Disease group and 14.7% of the Diabetes group were reassigned to a new group characterized by multiple cardiometabolic conditions.

CONCLUSIONS

The addition of Medicare data to VA data for older high-risk adults improved clustering model accuracy and altered the clinical profiles of groups. Accessing or accounting for multi-system data is key to the success of interventions based on empiric grouping in populations with dual-system use.

Electronic Health Literacy Among Baby Boomers: A Typology

BACKGROUND

Forecasts suggest that older adults will place unprecedented demands on future health care systems. Electronic health (eHealth) resources can potentially mitigate some pressures, but to be effective patients need to be able to use them. The negative relationship between eHealth literacy and age usually results in older adults classified as one homogenous mass, which misses the opportunity to tailor interventions.

OBJECTIVE

This research examines similarities and differences within the baby boom cohort among a sample that uses the internet for health information.

METHODS

We used an electronic survey with random samples of baby boomers (N = 996) from the United States, the United Kingdom, and New Zealand.

KEY RESULTS

Four distinct subgroups, or segments, emerged. While not different from a socioeconomic perspective, these four groups have very different levels of eHealth literacy and corresponding health behaviors. Therefore, we contribute a more complex picture than is usually presented in eHealth studies.

CONCLUSIONS

Resulting insights offer a useful starting point for providers wishing to better tailor health products, services, and communications to this large cohort of future older individuals. [HLRP: Health Literacy Research and Practice. 2024;8(1):e3-e11.].

Assigning disease clusters to people: A cohort study of the implications for understanding health outcomes in people with multiple long-term conditions

Background

Identifying clusters of co-occurring diseases may help characterise distinct phenotypes of Multiple Long-Term Conditions (MLTC). Understanding the associations of disease clusters with health-related outcomes requires a strategy to assign clusters to people, but it is unclear how the performance of strategies compare.

Aims

First, to compare the performance of methods of assigning disease clusters to people at explaining mortality, emergency department attendances and hospital admissions over one year. Second, to identify the extent of variation in the associations with each outcome between and within clusters.

Methods

We conducted a cohort study of primary care electronic health records in England, including adults with MLTC. Seven strategies were tested to assign patients to fifteen disease clusters representing 212 LTCs, identified from our previous work. We tested the performance of each strategy at explaining associations with the three outcomes over 1 year using logistic regression and compared to a strategy using the individual LTCs.

Results

6,286,233 patients with MLTC were included. Of the seven strategies tested, a strategy assigning the count of conditions within each cluster performed best at explaining all three outcomes but was inferior to using information on the individual LTCs. There was a larger range of effect sizes for the individual LTCs within the same cluster than there was between the clusters.

Conclusion

Strategies of assigning clusters of co-occurring diseases to people were less effective at explaining health-related outcomes than a person's individual diseases. Furthermore, clusters did not represent consistent relationships of the LTCs within them, which might limit their application in clinical research.

Supporting professionals to implement integrated, person-centered care for people with chronic conditions: the TARGET pilot study

OBJECTIVE

The TARGET program for integrated, person-centered care for people with chronic conditions offers primary care (PC) professionals a set of tools and trainings to actively engage in population segmentation and person-centered needs assessments (PCNAs). A pilot study was conducted to gain insight into the program's feasibility and acceptability, and identify preconditions for successful implementation.

DESIGN AND SETTING

Seven Dutch PC practices participated in a half-year pilot study starting in August 2020. We performed a review of the population segmentation tool, observed four training sessions and 15 PCNAs, and interviewed 15 professionals and 12 patients.

RESULTS

Regarding feasibility and acceptability, we found that the tools and trainings provided professionals with skills to use the segmentation tool and take a more coaching role in the well-appreciated PCNAs. Concerning implementation preconditions, we found that team commitment and network connections need improvement, although work pleasure increased and professionals generally wanted the program to continue.

CONCLUSIONS

While the content of the TARGET program is supported by its users, the implementation process, for instance team commitment to the program, needs more attention in future upscaling efforts.

How to extend value-based healthcare to population-based healthcare systems? Defining an outcome-based segmentation model for health authority

OBJECTIVES

Value-based healthcare (VBHC) is considered the most promising guiding principle for a new generation of health service production. Many countries have attempted to apply VBHC to managerial and clinical decision-making. However, implementation remains in its infancy and varies between countries. The objective of the study is to help health systems implement a value-based approach by building an outcome-based population segmentation model for health authorities (HAs).

DESIGN

First, we define the principles according to which segmentation models in healthcare could be developed. Second, we merge the theoretical characteristics of outcomes with population segmentation dimensions identified in previous literature and design a flow model that establishes population segments from these combinations. We then estimate the size of the segments based on national register data.

RESULTS

The population can be divided into 10 different segments based on relevant outcomes, goals and the outcome measurement logic. These segments consist of healthy, help, increased risk, mild curable without risk, mild curable with risk, severe curable without risk, severe curable with risk, single chronic, multimorbid and terminal. The representatives of Finnish HAs found the segments meaningful for evaluating and managing the healthcare system towards improved population health.

CONCLUSIONS

An outcome-based segmentation model for the entire population is needed if an HA wants to steer the healthcare system employing the principles of VBHC. Segmentation should be based on the outcome measurement logic and outcome measurements relevant to each segment and the number of segments has to be limited.

The impact of atypical intrahospital transfers on patient outcomes: a mixed methods study

The architectural design of hospitals worldwide is centred around individual departments, which require the movement of patients between wards. However, patients do not always take the simplest route from admission to discharge, but can experience convoluted movement patterns, particularly when bed availability is low. Few studies have explored the impact of these rarer, atypical trajectories. Using a mixed-method explanatory sequential study design, we firstly used three continuous years of electronic health record data prior to the Covid-19 pandemic, from 55,152 patients admitted to a London hospital network to define the ward specialities by patient type using the Herfindahl-Hirschman index. We explored the impact of 'regular transfers' between pairs of wards with shared specialities, 'atypical transfers' between pairs of wards with no shared specialities and 'site transfers' between pairs of wards in different hospital site locations, on length of stay, 30-day readmission and mortality. Secondly, to understand the possible reasons behind atypical transfers we conducted three focus groups and three in-depth interviews with site nurse practitioners and bed managers within the same hospital network. We found that at least one atypical transfer was experienced by 12.9% of patients. Each atypical transfer is associated with a larger increase in length of stay, 2.84 days (95% CI 2.56-3.12), compared to regular transfers, 1.92 days (95% CI 1.82-2.03). No association was found between odds of mortality, or 30-day readmission and atypical transfers after adjusting for confounders. Atypical transfers appear to be driven by complex patient conditions, a lack of hospital capacity, the need to reach specific services and facilities, and more exceptionally, rare events such as major incidents. Our work provides an important first step in identifying unusual patient movement and its impacts on key patient outcomes using a system-wide, data-driven approach. The broader impact of moving patients between hospital wards, and possible downstream effects should be considered in hospital policy and service planning.

Segmenting Patients With Diabetes With the Navigator Service in Primary Care and a Description of the Self-Acting Patient Group: Cross-Sectional Study

BACKGROUND

The aim of patient segmentation is to recognize patients with similar health care needs. The Finnish patient segmentation service Navigator segregates patients into 4 groups, including a self-acting group, who presumably manages their everyday life and coordinates their health care. Digital services could support their self-care. Knowledge on self-acting patients' characteristics is lacking.

OBJECTIVE

The study aims are to describe how Navigator assigns patients with diabetes to the 4 groups at nurses' appointments at a health center, the self-acting patient group's characteristics compared with other patient groups, and the concordance between the nurse's evaluation of the patient's group and the actual group assigned by Navigator (criterion validity).

METHODS

Patients with diabetes ≥18 years old visiting primary care were invited to participate in this cross-sectional study. Patients with disability preventing informed consent for participation were excluded. Nurses estimated the patients' upcoming group results before the appointment. We describe the concordance (%) between the evaluation and actual groups. Nurses used Navigator patients with diabetes (n=304) at their annual follow-up visits. The self-acting patients' diabetes care values (glycated hemoglobin [HbA1c], urine albumin to creatinine ratio, low-density lipoprotein cholesterol, blood pressure, BMI), chronic conditions, medication, smoking status, self-rated health, disability (World Health Organization Disability Assessment Schedule [WHODAS] 2.0), health-related quality of life (EQ-5D-5L), and well-being (Well-being Questionnaire [WBQ-12]) and the patients' responses to Navigator's question concerning their digital skills as outcome variables were compared with those of the other patients. We used descriptive statistics for the patients' distribution into the 4 groups and demographic data. We used the Mann-Whitney U test with nonnormally distributed variables, independent samples t test with normally distributed variables, and Pearson chi-square tests with categorized variables to compare the groups.

RESULTS

Most patients (259/304, 85.2%) were in the self-acting group. Hypertension, hyperlipidemia, and joint ailments were the most prevalent comorbidities among all patients. Self-acting patients had less ischemic cardiac disease (P=.001), depression or anxiety (P=.03), asthma or chronic obstructive pulmonary disease (P<.001), long-term pain (P<.001), and related medication. Self-acting patients had better self-rated health (P<.001), functional ability (P<.001), health-related quality of life (P<.001), and general well-being (P<.001). All patients considered their skills at using electronic services to be good.

CONCLUSIONS

The patients in the self-acting group had several comorbidities. However, their functional ability was not yet diminished compared with patients in the other groups. Therefore, to prevent diabetic complications and disabilities, support for patients' self-management should be emphasized in their integrated care services. Digital services could be involved in the care of patients willing to use them. The study was performed in 1 health center, the participants were volunteers, and most patients were assigned to self-acting patient group. These facts limit the generalizability of our results.

INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID)

RR2-10.2196/20570.

The Shieh Score as a Risk Assessment Instrument for Reducing Hospital-Acquired Pressure Injuries: A Prospective Cohort Study

PURPOSE

The purpose of this study was to evaluate the Shieh Score's effectiveness in decreasing the rate of hospital-acquired pressure injuries when combined with an early warning notification system and standard order set of preventative measures.

DESIGN

This was a prospective cohort study.

SUBJECTS AND SETTING

This target population was nonpregnant, adult, hospitalized patients on inpatient and observation status at a tertiary hospital (Kaiser Permanente, Baldwin Park, California) during the 2020 year of the COVID-19 pandemic.

METHODS

A new, risk assessment instrument, the Shieh Score, was developed in 2019 to predict hospitalized patients at high risk for pressure injuries. Data collection occurred between January 21, 2020, and December 31, 2020. When a hospital patient met the high-risk criteria for the Shieh Score, a provider-ordered pink-colored sheet of paper titled "Skin at Risk" was hung at the head of the bed and a standard order set of pressure injury preventative measures was implemented by nursing staff.

RESULTS

Implementation of the program (Shieh Score, early warning system, and standard order set for preventive interventions) resulted in a 38% reduction in the annual hospital-acquired pressure injury rate from a mean incidence rate of 1.03 to 0.64 hospital-acquired pressure injuries per 1000 patient-days measured for the year 2020.

CONCLUSION

The Shieh Score is a pressure injury risk assessment instrument, which effectively identifies patients at high risk for hospital-acquired pressure injuries and decreases the hospital-acquired pressure injury rate when combined with an early warning notification system and standard order set.

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.

Are older people worse off in 2040 regarding health and resources to deal with it? - Future developments in complex health problems and in the availability of resources to manage health problems in the Netherlands

Introduction

Developing sustainable health policy requires an understanding of the future demand for health and social care. We explored the characteristics of the 65+ population in the Netherlands in 2020 and 2040, focusing on two factors that determine care needs: (1) the occurrence of complex health problems and (2) the availability of resources to manage health and care (e.g., health literacy, social support).

Methods

Estimations of the occurrence of complex health problems and the availability of resources for 2020 were based on registry data and patient-reported data. Estimations for 2040 were based on (a) expected demographic developments, and (b) expert opinions using a two-stage Delphi study with 26 experts from policy making, practice and research in the field of health and social care.

Results

The proportion of people aged 65+ with complex health problems and limited resources is expected to increase from 10% in 2020 to 12% in 2040 based on demographic developments, and to 22% in 2040 based on expert opinions. There was high consensus (>80%) that the proportion with complex health problems would be greater in 2040, and lower consensus (50%) on an increase of the proportion of those with limited resources. Developments that are expected to drive the future changes refer to changes in multimorbidity and in psychosocial status (e.g., more loneliness).

Conclusion

The expected increased proportion of people aged 65+ with complex health problems and limited resources together with the expected health and social care workforce shortages represent large challenges for public health and social care policy.

Applying a data-driven population segmentation approach in German claims data

BACKGROUND

Segmenting the population into homogenous groups according to their healthcare needs may help to understand the population's demand for healthcare services and thus support health systems to properly allocate healthcare resources and plan interventions. It may also help to reduce the fragmented provision of healthcare services. The aim of this study was to apply a data-driven utilisation-based cluster analysis to segment a defined population in the south of Germany.

METHODS

Based on claims data of one big German health insurance a two-stage clustering approach was applied to group the population into segments. A hierarchical method (Ward's linkage) was performed to determine the optimal number of clusters, followed by a k-means cluster analysis using age and healthcare utilisation data in 2019. The resulting segments were described in terms of their morbidity, costs and demographic characteristics.

RESULTS

The 126,046 patients were divided into six distinct population segments. Healthcare utilisation, morbidity and demographic characteristics differed significantly across the segments. The segment "High overall care use" comprised the smallest share of patients (2.03%) but accounted for 24.04% of total cost. The overall utilisation of services was higher than the population average. In contrast, the segment "Low overall care use" included 42.89% of the study population, accounting for 9.94% of total cost. Utilisation of services by patients in this segment was lower than population average.

CONCLUSION

Population segmentation offers the opportunity to identify patient groups with similar healthcare utilisation patterns, patient demographics and morbidity. Thereby, healthcare services could be tailored for groups of patients with similar healthcare needs.

Using the Johns Hopkins ACG Case-Mix System for population segmentation in a hospital-based adult patient population in Singapore

OBJECTIVE

Population health management involves risk characterisation and patient segmentation. Almost all population segmentation tools require comprehensive health information spanning the full care continuum. We assessed the utility of applying the ACG System as a population risk segmentation tool using only hospital data.

DESIGN

Retrospective cohort study.

SETTING

Tertiary hospital in central Singapore.

PARTICIPANTS

100 000 randomly selected adult patients from 1 January to 31 December 2017.

INTERVENTION

Hospital encounters, diagnoses codes and medications prescribed to the participants were used as input data to the ACG System.

PRIMARY AND SECONDARY OUTCOME MEASURES

Hospital costs, admission episodes and mortality of these patients in the subsequent year (2018) were used to assess the utility of ACG System outputs such as resource utilisation bands (RUBs) in stratifying patients and identifying high hospital care users.

RESULTS

Patients placed in higher RUBs had higher prospective (2018) healthcare costs, and were more likely to have healthcare costs in the top five percentile, to have three or more hospital admissions, and to die in the subsequent year. A combination of RUBs and ACG System generated rank probability of high healthcare costs, age and gender that had good discriminatory ability for all three outcomes, with area under the receiver-operator characteristic curve (AUC) values of 0.827, 0.889 and 0.876, respectively. Application of machine learning methods improved AUCs marginally by about 0.02 in predicting the top five percentile of healthcare costs and death in the subsequent year.

CONCLUSION

A population stratification and risk prediction tool can be used to appropriately segment populations in a hospital patient population even with incomplete clinical data.

Standardised Practice-Based Oral Health Data Collection: A Pilot Study in Different Countries

BACKGROUND

The Oral Health Observatory (OHO), launched in 2014 by FDI World Dental Federation, aims to provide a coordinated approach to international oral health data collection. A feasibility project involving 12 countries tested the implementation of the methodology and data collection tools and assessed data quality from 6 countries.

METHODS

National dental associations (NDAs) recruited dentists following a standardised sampling method. Dentists and patients completed paired questionnaires (N = 7907) about patients' demographics, dental attendance, oral health-related behaviours, oral impacts, and clinical measures using a mobile app. In addition, participating dentists (n = 93) completed an evaluation survey, and NDAs completed a survey and participated in workshops to assess implementation feasibility.

RESULTS

Feasibility data are presented from the 12 participating countries. In addition, the 6 countries most advanced with data collection as of July 2020 (China, Colombia, India, Italy, Japan, and Lebanon) were included in the assessment of data quality and qualitative evaluation of implementation feasibility. All NDAs in these 6 countries reported interest in collecting standardised, international data for policy and communication activities and to understand service use and needs. Eighty-two percent of dentists (n = 76) reported a patient response rate of between 80% and 100%. More than 70% (n = 71) of dentists were either satisfied or very satisfied with the patient recruitment and data collection methods. There were variations in patient oral health and behaviours across countries, such as self-reporting twice-daily brushing which ranged from 45% in India to 83% in Colombia.

CONCLUSIONS

OHO provides a feasible model for collecting international standardised data in dental practices. Reducing time implications, ensuring mobile app reliability, and allowing practitioners to access patient-reported outcomes to inform practice may enhance implementation.

A data-driven population health management approach: The Extramural LUMC Academic Network data infrastructure

Improving population health and reducing inequalities through better integrated health and social care services is high up on the agenda of policymakers internationally. In recent years, regional cross-domain partnerships have emerged in several countries, which aim to achieve better population health, quality of care and a reduction in the per capita costs. These cross-domain partnerships aim to have a strong data foundation and are committed to continuous learning in which data plays an essential role. This paper describes our approach towards the development of the regional integrative population-based data infrastructure Extramural LUMC (Leiden University Medical Center) Academic Network (ELAN), in which we linked routinely collected medical, social and public health data at the patient level from the greater The Hague and Leiden area. Furthermore, we discuss the methodological issues of routine care data and the lessons learned about privacy, legislation and reciprocities. The initiative presented in this paper is relevant for international researchers and policy-makers because a unique data infrastructure has been set up that contains data across different domains, providing insights into societal issues and scientific questions that are important for data driven population health management approaches.

Healthcare costs of diabetic foot disease in Italy: estimates for event and state costs

OBJECTIVE

This study aimed to estimate healthcare costs of diabetic foot disease (DFD) in a large population-based cohort of people with type-2 diabetes (T2D) in the Tuscany region (Italy).

DATA SOURCES/STUDY SETTING

Administrative healthcare data of Tuscany region, with 2018 as the base year.

STUDY DESIGN

Retrospective study assessing a longitudinal cohort of patients with T2D.

DATA COLLECTION/EXTRACTION METHODS

Using administrative healthcare data, DFD were identified using the International Classification of Diseases, Ninth Revision, Clinical Modification codes.

METHODS

We examined the annual healthcare costs of these clinical problems in patients with T2D between 2015 and 2018; moreover, we used a generalized linear model to estimate the total healthcare costs.

PRINCIPAL FINDINGS

Between 2015 and 2018, patients with T2D experiencing DFD showed significantly higher average direct costs than patients with T2D without DFD (p < 0.0001). Among patients with T2D experiencing DFD, those who experienced complications either in 2015-2017 and in 2018 incurred the highest incremental costs (incremental cost of € 16,702) followed by those with complications in 2018 only (incremental cost of € 9,536) and from 2015 to 2017 (incremental cost of € 800).

CONCLUSIONS

DFD significantly increase healthcare utilization and costs among patients with TD2. Healthcare costs of DFD among patients with T2D are associated with the timing and frequency of DFD. These findings should increase awareness among policymakers regarding resource reallocation toward preventive strategies among patients with T2D.

Service design in healthcare: a segmentation-based approach

Purpose

The study aims to explore how segmentation as a methodology can be adapted to the healthcare context to provide a more nuanced understanding of the served population and to facilitate the design of patient-centric services.

Design/methodology/approach

The study was based on a collaborative project with a national healthcare organization following the principles of action design research. The study describes the quantitative segmentation performed during the project, followed by a qualitative interview study of how segments correspond with patient behaviors in an actual healthcare setting, and service design workshops facilitated by segments. A number of design principles are outlined based on the learnings of the project.

Findings

The segmentation approach increased understanding of patient variability within the service provider organization and was considered an effective foundation for modular service design. Patient characteristics and life circumstances were related to specific patterns of health behaviors, such as avoidance or passivity, or a persistent proactivity. These patterns influenced the patients' preferred value co-creation role and what type of support patients sought from the care provider.

Practical implications

The proposed segmentation approach is immediately generalizable to further healthcare contexts and similar services: improved understanding of patients, vulnerable patients in particular, improves the fit and inclusivity of services.

Originality/value

The segmentation approach to service design was demonstrated to be effective in a large-scale context. The approach allows service providers to design service options that improve the fit with individual patients' needs for support and autonomy. The results illuminate how patient characteristics influence health and value co-creation behaviors.

Ambulatory Healthcare Use Profiles of Patients With Diabetes and Their Association With Quality of Care: A Cross-Sectional Study

BACKGROUND

Despite the growing burden of diabetes worldwide, evidence regarding the optimal models of care to improve the quality of diabetes care remains equivocal. This study aimed to identify profiles of patients with distinct ambulatory care use patterns and to examine the association of these profiles with the quality of diabetes care.

METHODS

We performed a cross-sectional study of the baseline data of 550 non-institutionalized adults included in a prospective, community-based, cohort study on diabetes care conducted in Switzerland. Clusters of participants with distinct patterns of ambulatory healthcare use were identified using discrete mixture models. To measure the quality of diabetes care, we used both processes of care indicators (eye and foot examination, microalbuminuria screening, blood cholesterol and glycated hemoglobin measurement [HbA1c], influenza immunization, blood pressure measurement, physical activity and diet advice) and outcome indicators (12-Item Short-Form Health Survey [SF-12], Audit of Diabetes-Dependent Quality of Life [ADDQoL], Patient Assessment of Chronic Illness Care [PACIC], Diabetes Self-Efficacy Scale, HbA1c value, and blood pressure <140/90 mmHg). For each profile of ambulatory healthcare use, we calculated adjusted probabilities of receiving processes of care and estimated adjusted outcomes of care using logistic and linear regression models, respectively.

RESULTS

Four profiles of ambulatory healthcare use were identified: participants with more visits to the general practitioner [GP] than to the diabetologist and receiving concomitant podiatry care ("GP & podiatrist", n=86); participants visiting almost exclusively their GP ("GP only", n=195); participants with a substantially higher use of all ambulatory services ("High users", n=96); and participants reporting more visits to the diabetologist and less visits to the GP than other profiles ("Diabetologist first", n=173). Whereas participants belonging to the "GP only" profile were less likely to report most processes related to the quality of diabetes care, outcomes of care were relatively comparable across all ambulatory healthcare use profiles.

CONCLUSIONS

Slight differences in quality of diabetes care appear across the four ambulatory healthcare use profiles identified in this study. Overall, however, results suggest that room for improvement exists in all profiles, and further investigation is necessary to determine whether individual characteristics (like diabetes-related factors) and/or healthcare factors contribute to the differences observed between profiles.

Patterns of healthcare utilisation in children and young people: a retrospective cohort study using routinely collected healthcare data in Northwest London

OBJECTIVES

With a growing role for health services in managing population health, there is a need for early identification of populations with high need. Segmentation approaches partition the population based on demographics, long-term conditions (LTCs) or healthcare utilisation but have mostly been applied to adults. Our study uses segmentation methods to distinguish patterns of healthcare utilisation in children and young people (CYP) and to explore predictors of segment membership.

DESIGN

A retrospective cohort study.

SETTING

Routinely collected primary and secondary healthcare data in Northwest London from the Discover database.

PARTICIPANTS

378 309 CYP aged 0-15 years registered to a general practice in Northwest London with 1 full year of follow-up.

PRIMARY AND SECONDARY OUTCOME MEASURES

Assignment of each participant to a segment defined by seven healthcare variables representing primary and secondary care attendances, and description of utilisation patterns by segment. Predictors of segment membership described by age, sex, ethnicity, deprivation and LTCs.

RESULTS

Participants were grouped into six segments based on healthcare utilisation. Three segments predominantly used primary care, two moderate utilisation segments differed in use of emergency or elective care, and a high utilisation segment, representing 16 632 (4.4%) children accounted for the highest mean presentations across all service types. The two smallest segments, representing 13.3% of the population, accounted for 62.5% of total costs. Younger age, residence in areas of higher deprivation and the presence of one or more LTCs were associated with membership of higher utilisation segments, but 75.0% of those in the highest utilisation segment had no LTC.

CONCLUSIONS

This article identifies six segments of healthcare utilisation in CYP and predictors of segment membership. Demographics and LTCs may not explain utilisation patterns as strongly as in adults, which may limit the use of routine data in predicting utilisation and suggest children have less well-defined trajectories of service use than adults.

Digital health opportunities to improve Primary Health Care in the context of COVID-19: A Scoping Review (Preprint)

Background

The COVID-19 pandemic brought social, economic, and health impacts, requiring fast adaptation of health systems. Although information and communication technologies were essential for achieving this objective, the extent to which health systems incorporated this technology is unknown.

Objective

The aim of this study was to map the use of digital health strategies in primary health care worldwide and their impact on quality of care during the COVID-19 pandemic.

Methods

We performed a scoping review based on the Joanna Briggs Institute manual and guided by the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) Extension for Scoping Reviews. A systematic and comprehensive three-step search was performed in June and July 2021 in multidisciplinary health science databases and the gray literature. Data extraction and eligibility were performed by two authors independently and interpreted using thematic analysis.

Results

A total of 44 studies were included and six thematic groups were identified: characterization and geographic distribution of studies; nomenclatures of digital strategies adopted; types of information and communication technologies; characteristics of digital strategies in primary health care; impacts on quality of care; and benefits, limitations, and challenges of digital strategies in primary health care. The impacts on organization of quality of care were investigated by the majority of studies, demonstrating the strengthening of (1) continuity of care; (2) economic, social, geographical, time, and cultural accessibility; (3) coordination of care; (4) access; (5) integrality of care; (6) optimization of appointment time; (7) and efficiency. Negative impacts were also observed in the same dimensions, such as reduced access to services and increased inequity and unequal use of services offered, digital exclusion of part of the population, lack of planning for defining the role of professionals, disarticulation of actions with real needs of the population, fragile articulation between remote and face-to-face modalities, and unpreparedness of professionals to meet demands using digital technologies.

Conclusions

The results showed the positive and negative impacts of remote strategies on quality of care in primary care and the inability to take advantage of the potential of technologies. This may demonstrate differences in the organization of fast and urgent implementation of digital strategies in primary health care worldwide. Primary health care must strengthen its response capacity, expand the use of information and communication technologies, and manage challenges using scientific evidence since digital health is important and must be integrated into public service.

Acceptance and Perceived Control are Independently Associated With Quality of Life in Inflammatory Bowel Disease: Introduction of a New Segmentation Model

BACKGROUND AND AIMS

Segmentation of patients based on psychological determinants of subjective health may provide new ways to personalized care. The cross-disease segmentation model developed by Bloem & Stalpers discriminates patients based on disease acceptance and perceived control. We aimed to validate the segmentation model, compare segments and evaluate whether segments independently correlate with quality of life in inflammatory bowel disease [IBD].

METHODS

A cross-sectional study of adult IBD patients was performed with questionnaires on quality of life [32-item inflammatory bowel disease questionnaire], acceptance and perceived control [six items with 7-point Likert scale]. Four segments were formed [cut-off > 5]: [I] high acceptance, high control; [II] high acceptance, low control [III]; low acceptance, high control and; [IV] low acceptance, low control.

RESULTS

We included 686 patients. The acceptance and perceived control scales were unidimensionally structured and internally consistent. Segments differed significantly in age, smoking behaviour, diagnosis, disease duration, extra-intestinal manifestations, IBD medication, clinical disease activity and quality of life. High acceptance (standardized beta coefficient [ß] 0.25, p < 0.001), high perceived control [ß 0.12, p < 0.001] or both [ß 0.53, p < 0.001], were associated with a significantly better health-related quality of life compared with low acceptance and low perceived control. Sociodemographic and clinical factors explained 25% of the variance in quality of life. The explained variance significantly increased to 45% when the patients' segment was added to the model [ΔR2 20%, p < 0.001].

CONCLUSIONS

The segmentation model based on disease acceptance and perceived control is valid in IBD patients and discriminates different segments that correlate independently with quality of life. This may open new strategies for patient care.

Improving Physical Health Among People With Serious Mental Illness: The Role of the Specialty Mental Health Sector

People with serious mental illness die 10-20 years earlier, compared with the overall population, and the excess mortality is driven by undertreated physical health conditions. In the United States, there is growing interest in models integrating physical health care delivery, management, or coordination into specialty mental health programs, sometimes called "reverse integration." In November 2019, the Johns Hopkins ALACRITY Center for Health and Longevity in Mental Illness convened a forum of 25 experts to discuss the current state of the evidence on integrated care models based in the specialty mental health system and to identify priorities for future research, policy, and practice. This article summarizes the group's conclusions. Key research priorities include identifying the active ingredients in multicomponent integrated care models and developing and validating integration performance metrics. Key policy and practice recommendations include developing new financing mechanisms and implementing strategies to build workforce and data capacity. Forum participants also highlighted an overarching need to address socioeconomic risks contributing to excess mortality among adults with serious mental illness.

Early Implementation of a Patient-Centered Medical Home in Singapore: A Qualitative Study Using Theory on Diffusion of Innovations

Patient-Centered Medical Home (PCMH) has been found to improve care for complex needs patients in some countries but has not yet been widely adopted in Singapore. This study explored the ground-up implementation of a PCMH in Singapore by describing change strategies and unpacking initial experience and perception. In-depth interviews were conducted for twenty-two key informants from three groups: the implementers, their implementation partners, and other providers. “Diffusion of innovations” emerged as an overarching theory to contextualize PCMH in its early implementation. Three core “innovations” differentiated the PCMH from usual primary care: (i) team-based and integrated care; (ii) empanelment; and (iii) shared care with other general practitioners. Change strategies employed to implement these innovations included repurposing pre-existing resources, building a partnership to create supporting infrastructure and pathways in the delivery system, and doing targeted outreach to introduce the PCMH. Initial experience and perception were characterized by processes to “adopt” and “assimilate” the innovations, which were identified as challenging due to less predictable, self-organizing behaviors by multiple players. To work with the inherent complexity and novelty of the innovations, time, leadership, standardized methods, direct communication, and awareness-building efforts are needed. This study was retrospectively registered (Protocol ID: NCT04594967).

Using social marketing for the promotion of cognitive health: a scoping review protocol

INTRODUCTION

The use of social marketing strategies to induce the promotion of cognitive health has received little attention in research. The objective of this scoping review is twofold: (i) to identify the social marketing strategies that have been used in recent years to initiate and maintain health-promoting behaviour; (ii) to advance research in this area to inform policy and practice on how to best make use of these strategies to promote cognitive health.

METHODS AND ANALYSIS

We will use the five-stage methodological framework of Arksey and O'Malley. Articles in English published since 2010 will be searched in electronic databases (the Cochrane Library, DoPHER, the International Bibliography of the Social Sciences, PsycInfo, PubMed, ScienceDirect, Scopus). Quantitative and qualitative study designs as well as reviews will be considered. We will include those articles that report the design, implementation, outcomes and evaluation of programmes and interventions concerning social marketing and/or health promotion and/or promotion of cognitive health. Grey literature will not be searched. Two independent reviewers will assess in detail the abstracts and full text of selected citations against the inclusion criteria. A Preferred Reporting Items for Systematic Reviews and Meta-Analyses flowchart for Scoping Reviews will be used to illustrate the process of article selection. We will use a data extraction form, present the results through narrative synthesis and discuss them in relation to the scoping review research questions.

ETHICS AND DISSEMINATION

Ethics approval is not required for conducting this scoping review. The results of the review will be the first step to advance a conceptual framework, which contributes to the development of interventions targeting the promotion of cognitive health. The results will be published in a peer-reviewed scientific journal. They will also be disseminated to key stakeholders in the field of the promotion of cognitive health.

Socially situated risk: challenges and strategies for implementing algorithmic risk scoring for care management

OBJECTIVE

To characterize challenges and strategies related to algorithmic risk scoring for care management eligibility determinations.

MATERIALS AND METHODS

Interviews with 19 administrators from 13 physician organizations representing over 2200 physician offices and 8800 physicians in Michigan. Post-implementation interviews were coded using thematic analysis.

RESULTS

Utility of algorithmic risk scores was limited due to outdated claims or incomplete information about patients' socially situated risks (eg, caregiver turnover, social isolation). Resulting challenges included lack of physician engagement and inefficient use of staff time reviewing eligibility determinations. To address these challenges, risk scores were supplemented with physician knowledge and clinical data.

DISCUSSION AND CONCLUSION

Current approaches to risk scoring based on claims data for payer-led programs struggle to gain physician acceptance and support because of data limitations. To respond to these limitations, physician input regarding socially situated risk and utilization of more timely data may improve eligibility determinations.

Developing High Performance Secure Multi-Party Computation Protocols in Healthcare: A Case Study of Patient Risk Stratification

We demonstrate that secure multi-party computation (MPC) using garbled circuits is viable technology for solving clinical use cases that require cross-institution data exchange and collaboration. We describe two MPC protocols, based on Yao's garbled circuits and tested using large and realistically synthesized datasets. Linking records using private set intersection (PSI), we compute two metrics often used in patient risk stratification: high utilizer identification (PSI-HU) and comorbidity index calculation (PSI-CI). Cuckoo hashing enables our protocols to achieve extremely fast run times, with answers to clinically meaningful questions produced in minutes instead of hours. Also, our protocols are provably secure against any computationally bounded adversary in a semi-honest setting, the de-facto mode for cross-institution data analytics. Finally, these protocols eliminate the need for an implicitly trusted third-party "honest broker" to mediate the information linkage and exchange.

Population segmentation of type 2 diabetes mellitus patients and its clinical applications - a scoping review

Background

Population segmentation permits the division of a heterogeneous population into relatively homogenous subgroups. This scoping review aims to summarize the clinical applications of data driven and expert driven population segmentation among Type 2 diabetes mellitus (T2DM) patients.

Methods

The literature search was conducted in Medline®, Embase®, Scopus® and PsycInfo®. Articles which utilized expert-based or data-driven population segmentation methodologies for evaluation of outcomes among T2DM patients were included. Population segmentation variables were grouped into five domains (socio-demographic, diabetes related, non-diabetes medical related, psychiatric / psychological and health system related variables). A framework for PopulAtion Segmentation Study design for T2DM patients (PASS-T2DM) was proposed.

Results

Of 155,124 articles screened, 148 articles were included. Expert driven population segmentation approach was most commonly used, of which judgemental splitting was the main strategy employed (n = 111, 75.0%). Cluster based analyses (n = 37, 25.0%) was the main data driven population segmentation strategies utilized. Socio-demographic (n = 66, 44.6%), diabetes related (n = 54, 36.5%) and non-diabetes medical related (n = 18, 12.2%) were the most used domains. Specifically, patients’ race, age, Hba1c related parameters and depression / anxiety related variables were most frequently used. Health grouping/profiling (n = 71, 48%), assessment of diabetes related complications (n = 57, 38.5%) and non-diabetes metabolic derangements (n = 42, 28.4%) were the most frequent population segmentation objectives of the studies.

Conclusions

Population segmentation has a wide range of clinical applications for evaluating clinical outcomes among T2DM patients. More studies are required to identify the optimal set of population segmentation framework for T2DM patients.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12874-021-01209-w.

Population Segmentation Based on Healthcare Needs: Validation of a Brief Clinician-Administered Tool

BACKGROUND

As populations age with increasingly complex chronic conditions, segmenting populations into clinically meaningful categories of healthcare and related service needs can provide healthcare planners with crucial information to optimally meet needs. However, while conventional approaches typically involve electronic medical records (EMRs), such records do not always capture information reliably or accurately.

OBJECTIVE

We describe the inter-rater reliability and predictive validity of a clinician-administered tool, the Simple Segmentation Tool (SST) for categorizing older individuals into one of six Global Impression (GI) segments and eight complicating factors (CFs) indicative of healthcare and related social needs.

DESIGN

Observational study ( ClinicalTrials.gov , number NCT02663037).

PARTICIPANTS

Patients aged 55 years and above.

MAIN MEASURES

Emergency department (ED) subjects (between May and June 2016) had baseline SST assessment by two physicians and a nurse concurrently seeing the same individual. General medical (GM) ward subjects (February 2017) had a SST assessment by their principal physician. Adverse events (ED visits, hospitalizations, and mortality over 90 days from baseline) were determined by a blinded reviewer. Inter-rater reliability was measured using Cohen's kappa. Predictive validity was evaluated using Cox hazard ratios based on time to first adverse event.

KEY RESULTS

Cohen's kappa between physician-physician, service physician-nurse, and physician-nurse pairs for GI were 0.60, 0.71, and 0.68, respectively. Cox analyses demonstrated significant predictive validity of GI and CFs for adverse outcomes.

CONCLUSIONS

With modest training, clinicians can complete a brief instrument to segment their patient into clinically meaningful categories of healthcare and related service needs. This approach can complement and overcome current limitations of EMR-based instruments, particularly with respect to whole-patient care.

TRIAL REGISTRATION

ClinicalTrials.gov Identifier: NCT02663037.

A Value-Based Steering Model for Healthcare

This article aims to answer how a commissioning body can steer health services based on value in an environment where the commissioner is responsible for the health services of a population with varying health service needs. In this design science study, we constructed a value-based steering model consisting of three parts: (1) the principles of steering; (2) the steering process; and (3) Value Steering Canvas, a concrete tool for steering. The study is based on Finland, a tax-funded healthcare system, where healthcare is a public service. The results can be applied in any system where there is a commissioner and a service provider, whether they are two separate organizations or not. We conclude that steering can be done based on value. The commissioning body can start using value-based steering without changes in legislation or in the present service system. Further research is needed to test the model in practice.

Web-Based Patient Segmentation in Finnish Primary Care: Protocol for Clinical Validation of the Navigator Service in Patients With Diabetes

BACKGROUND

An aging population and increasing multimorbidity challenge health care systems worldwide. Patient segmentation aims to recognize groups of patients with similar needs, offer targeted services to these groups, and reduce the burden of health care. In this study, the unique Finnish innovation Navigator, a web-based service for patient segmentation, is presented. Both patients and health care professionals complete the electronic questionnaire concerning patients' coping in everyday life and health state. Thus, it considers the patient perspective on self-care. One of four customership-strategy (CS) groups (self-acting, community, cooperating, and network) is then proposed in response to the answers given. This resulting strategy helps both professionals to coordinate patient health care and patients to utilize appropriate health services.

OBJECTIVE

This study aims to determine the feasibility, validity, and reliability of the Navigator service in the segmentation of patients with diabetes into four CS groups in a primary care setting. Patient characteristics concerning demographic status, chronic conditions, disabilities, health-related quality of life, and well-being in different CS groups will be described. We hypothesize that patients in the network group will be older, have more illnesses, chronic conditions or disabilities, and require more health care services than patients in the self-acting group.

METHODS

In this mixed methods study, data collection was based on questionnaires (user experience of Navigator, demographic and health status, World Health Organization Disability Assessment Schedule 2.0, EuroQol 5D, Wellbeing Questionnaire 12, and the Diabetes Treatment Satisfaction Questionnaire) issued to 300 patients with diabetes and on user-experience questionnaires for and semistructured focus-group interviews with 12 nurses. Navigator-database reports and diabetes-care values (blood pressure, BMI, HbA1c, low-density lipoprotein, albumin-creatinine, smoking status) were collected. Qualitative and descriptive analyses were used to study the feasibility, content, concurrent, and face validity of Navigator. While criterion and concurrent validity were examined with correlations, reliability was examined by calculating Cohen kappa and Cronbach alpha. Construct validity is studied by performing exploratory-factor analysis on Navigator data reports and by hypothesis testing. The values, demographics, and health status of patients in different groups were described, and differences between groups were studied by comparing means. Linear regression analysis was performed to assess which variables affect CS group variation.

RESULTS

Data collection was completed in September 2019, and the first feasibility results are expected by the end of 2020. Further results and publications are expected in 2021 and 2022.

CONCLUSIONS

This is the first scientific study concerning Navigator's psychometric properties. The study will examine the segregation of patients with diabetes into four CS groups in a primary care setting and the differences between patients in groups. This study will assist in Navigator's further development as a patient segmentation method considering patients' perspectives on self-care. This study will not prove the effectiveness or efficacy of Navigator; therefore, it is essential to study these outcomes of separate care pathways.

INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID)

DERR1-10.2196/20570.

Population Health Management to identify and characterise ongoing health need for high-risk individuals shielded from COVID-19: a cross-sectional cohort study

OBJECTIVES

To use Population Health Management (PHM) methods to identify and characterise individuals at high-risk of severe COVID-19 for which shielding is required, for the purposes of managing ongoing health needs and mitigating potential shielding-induced harm.

DESIGN

Individuals at 'high risk' of COVID-19 were identified using the published national 'Shielded Patient List' criteria. Individual-level information, including current chronic conditions, historical healthcare utilisation and demographic and socioeconomic status, was used for descriptive analyses of this group using PHM methods. Segmentation used k-prototypes cluster analysis.

SETTING

A major healthcare system in the South West of England, for which linked primary, secondary, community and mental health data are available in a system-wide dataset. The study was performed at a time considered to be relatively early in the COVID-19 pandemic in the UK.

PARTICIPANTS

1 013 940 individuals from 78 contributing general practices.

RESULTS

Compared with the groups considered at 'low' and 'moderate' risk (ie, eligible for the annual influenza vaccination), individuals at high risk were older (median age: 68 years (IQR: 55-77 years), cf 30 years (18-44 years) and 63 years (38-73 years), respectively), with more primary care/community contacts in the previous year (median contacts: 5 (2-10), cf 0 (0-2) and 2 (0-5)) and had a higher burden of comorbidity (median Charlson Score: 4 (3-6), cf 0 (0-0) and 2 (1-4)). Geospatial analyses revealed that 3.3% of rural and semi-rural residents were in the high-risk group compared with 2.91% of urban and inner-city residents (p<0.001). Segmentation uncovered six distinct clusters comprising the high-risk population, with key differentiation based on age and the presence of cancer, respiratory, and mental health conditions.

CONCLUSIONS

PHM methods are useful in characterising the needs of individuals requiring shielding. Segmentation of the high-risk population identified groups with distinct characteristics that may benefit from a more tailored response from health and care providers and policy-makers.

Tailoring integrated care services for high-risk patients with multiple chronic conditions: a risk stratification approach using cluster analysis

Background

The purpose of this study was to produce a risk stratification within a population of high-risk patients with multiple chronic conditions who are currently treated under a case management program and to explore the existence of different risk subgroups. Different care strategies were then suggested for healthcare reform according to the characteristics of each subgroup.

Methods

All high-risk multimorbid patients from a case management program in the Navarra region of Spain were included in the study (n = 885). A 1-year mortality risk score was estimated for each patient by logistic regression. The population was then divided into subgroups according to the patients’ estimated risk scores. We used cluster analysis to produce the stratification with Ward’s linkage hierarchical algorithm. The characteristics of the resulting subgroups were analyzed, and post hoc pairwise tests were performed.

Results

Three distinct risk strata were found, containing 45, 38 and 17% of patients. Age increased from cluster to cluster, and functional status, clinical severity, nursing needs and nutritional values deteriorated. Patients in cluster 1 had lower renal deterioration values, and patients in cluster 3 had higher rates of pressure skin ulcers, higher rates of cerebrovascular disease and dementia, and lower prevalence rates of chronic obstructive pulmonary disease.

Conclusions

This study demonstrates the existence of distinct subgroups within a population of high-risk patients with multiple chronic conditions. Current case management integrated care programs use a uniform treatment strategy for patients who have diverse needs. Alternative treatment strategies should be considered to fit the needs of each patient subgroup.

Segmentation of health-care consumers: psychological determinants of subjective health and other person-related variables

Background

There is an observable, growing trend toward tailoring support programs – in addition to medical treatment – more closely to individuals to help improve patients’ health status. The segmentation model developed by Bloem & Stalpers [Nyenrode Research Papers Series 12:1–22, 2012] may serve as a solid basis for such an approach. The model is focused on individuals’ ‘health experience’ and is therefore a ‘cross-disease’ model. The model is based on the main psychological determinants of subjective health: acceptance and perceived control. The model identifies four segments of health-care consumers, based on high or low values on these determinants. The goal of the present study is twofold: the identification of criteria for differentiating between segments, and profiling of the segments in terms of socio-demographic and socio-economic variables.

Methods

The data (acceptance, perceived control, socio-economic, and socio-demographic variables) for this study were obtained by using an online survey (a questionnaire design), that was given (random sample N = 2500) to a large panel of Dutch citizens. The final sample consisted of 2465 participants – age distribution and education level distribution in the sample were similar to those in the Dutch population; there was an overrepresentation of females. To analyze the data factor analyses, reliability tests, descriptive statistics and t-tests were used.

Results

Cut-off scores, criteria to differentiate between the segments, were defined as the medians of the distributions of control and acceptance. Based on the outcomes, unique profiles have been formed for the four segments: 1. ‘Importance of self-management’ – relatively young, high social class, support programs: high-quality information. 2. ‘Importance of personal control’ – relatively old, living in rural areas, high in homeownership; supportive programs: developing personal control skills. 3. ‘Importance of acceptance’ – relatively young male; supportive programs: help by physicians and nurses. 4. ‘Importance of perspective and direction’ – female, low social class, receiving informal care; support programs: counseling and personal care.

Conclusions

The profiles describe four segments of individuals/patients that are clearly distinct from each other, each with its own description. The enriched descriptions provide a better basis for the allocation and developing of supportive programs and interventions across individuals.

Patient Segmentation: Adjust the Production Logic to the Medical Knowledge Applied and the Patient's Ability to Self-Manage-A Discussion Paper

This discussion paper argues that population segmentation according to healthcare needs and risks—the usual approach—might help to identify patients for targeted action, but does not inform how to design efficient service delivery. In other service industries customer segmentation is typically done based on customer preferences. Products or services are customized and marketing strategies designed to reach the most profitable customers and improve revenue generation. This paper presents an alternative approach, in which patient needs are matched with a production logic derived from the medical knowledge needed to manage the health problem, and patients' willingness and ability to self-manage and co-produce services. Seven segments are identified: healthy persons; persons with incidental needs; persons with chronic conditions; persons with multiple health problems and illnesses (often elderly); persons needing precise elective interventions; persons needing qualified accident and emergency services; and tertiary care patients. Designing care to suit these patient segments will use resources more efficiently, with better prospects of favorable medical outcomes, a higher service quality, less complications, and improved patient safety.

The development of a brief measure of health personality

The purpose of this research was to develop a brief assessment of health personality, defined as a set of individual dispositions that are directly related to health. In Study 1, an initial pool of items was developed with 615 older adults, 65 years of age and older. The scale was reduced to a 15-item version for use in applied health care settings. Results indicated that the 'Health Personality Assessment scale' has good internal consistency, and the five-factors correlated significantly with self-reported measures of physical health and well-being. In Study 2, the scale was cross-validated with 254 older adults from the Health Literacy and Cognitive Function among Older Adults Study. The scale was refined and a third study consisted of 3,907 older adults. Reliability and validity of the scale were confirmed. Future research should evaluate the usefulness of this scale in applied healthcare settings.

Do healthcare needs-based population segments predict outcomes among the elderly? Findings from a prospective cohort study in an urbanized low-income community

Background

A rapidly ageing population with increasing prevalence of chronic disease presents policymakers the urgent task of tailoring healthcare services to optimally meet changing needs. While healthcare needs-based segmentation is a promising approach to efficiently assessing and responding to healthcare needs at the population level, it is not clear how available schemes perform in the context of community-based surveys administered by non-medically trained personnel. The aim of this prospective cohort, community setting study is to evaluate 4 segmentation schemes in terms of practicality and predictive validity for future health outcomes and service utilization.

Methods

A cohort was identified from a cross-sectional health and social characteristics survey of Singapore public rental housing residents aged 60 years and above. Baseline survey data was used to assign individuals into segments as defined by 4 predefined population segmentation schemes developed in Singapore, Delaware, Lombardy and North-West London. From electronic data records, mortality, hospital admissions, emergency department visits, and specialist outpatient clinic visits were assessed for 180 days after baseline segment assignment and compared to segment membership for each segmentation scheme.

Results

Of 1324 residents contacted, 928 agreed to participate in the survey (70% response). All subjects could be assigned an exclusive segment for each segmentation scheme. Individuals in more severe segments tended to have lower quality of life as assessed by the EQ-5D Index for health utility. All population segmentation schemes were observed to exhibit an ability to differentiate different levels of mortality and healthcare utilization.

Conclusions

It is practical to assign individuals to healthcare needs-based population segments through community surveys by non-medically trained personnel. The resulting segments for all 4 schemes evaluated in this way have an ability to predict health outcomes and utilization over the medium term (180 days), with significant overlap for some segments. Healthcare needs-based segmentation schemes which are designed to guide action hold particular promise for promoting efficient allocation of services to meet the needs of salient population groups. Further evaluation is needed to determine if these schemes also predict responsiveness to interventions to meet needs implied by segment membership.

Identifying optimal indicators and purposes of population segmentation through engagement of key stakeholders: a qualitative study

BACKGROUND

Various population segmentation tools have been developed to inform the design of interventions that improve population health. However, there has been little consensus on the core indicators and purposes of population segmentation. The existing frameworks were further limited by their applicability in different practice settings involving stakeholders at all levels. The aim of this study was to generate a comprehensive set of indicators and purposes of population segmentation based on the experience and perspectives of key stakeholders involved in population health.

METHODS

We conducted in-depth semi-structured interviews using purposive sampling with key stakeholders (e.g. government officials, healthcare professionals, social service providers, researchers) involved in population health at three distinct levels (micro, meso, macro) in Singapore. The interviews were audio-recorded and transcribed verbatim. Thematic content analysis was undertaken using NVivo 12.

RESULTS

A total of 25 interviews were conducted. Eight core indicators (demographic characteristics, economic characteristics, behavioural characteristics, disease state, functional status, organisation of care, psychosocial factors and service needs of patients) and 21 sub-indicators were identified. Age and financial status were commonly stated as important indicators that could potentially be used for population segmentation across three levels of participants. Six intended purposes for population segmentation included improving health outcomes, planning for resource allocation, optimising healthcare utilisation, enhancing psychosocial and behavioural outcomes, strengthening preventive efforts and driving policy changes. There was consensus that planning for resource allocation and improving health outcomes were considered two of the most important purposes for population segmentation.

CONCLUSIONS

Our findings shed light on the need for a more person-centric population segmentation framework that incorporates upstream and holistic indicators to be able to measure population health outcomes and to plan for appropriate resource allocation. Core elements of the framework may apply to other healthcare settings and systems responsible for improving population health.

TRIAL REGISTRATION

The study was approved by the SingHealth Institutional Review Board (CIRB Reference number: 2017/2597).