Individualised risk estimation and the nature of prevention

Individual risk prediction: Comparing random forests with Cox proportional-hazards model by a simulation study

With big data becoming widely available in healthcare, machine learning algorithms such as random forest (RF) that ignores time-to-event information and random survival forest (RSF) that handles right-censored data are used for individual risk prediction alternatively to the Cox proportional hazards (Cox-PH) model. We aimed to systematically compare RF and RSF with Cox-PH. RSF with three split criteria [log-rank (RSF-LR), log-rank score (RSF-LRS), maximally selected rank statistics (RSF-MSR)]; RF, Cox-PH, and Cox-PH with splines (Cox-S) were evaluated through a simulation study based on real data. One hundred eighty scenarios were investigated assuming different associations between the predictors and the outcome (linear/linear and interactions/nonlinear/nonlinear and interactions), training sample sizes (500/1000/5000), censoring rates (50%/75%/93%), hazard functions (increasing/decreasing/constant), and number of predictors (seven, 15 including noise variables). Methods' performance was evaluated with time-dependent area under curve and integrated Brier score. In all scenarios, RF had the worst performance. In scenarios with a low number of events (⩽70), Cox-PH was at least noninferior to RSF, whereas under linearity assumption it outperformed RSF. Under the presence of interactions, RSF performed better than Cox-PH as the number of events increased whereas Cox-S reached at least similar performance with RSF under nonlinear effects. RSF-LRS performed slightly worse than RSF-LR and RSF-MSR when including noise variables and interaction effects. When applied to real data, models incorporating survival time performed better. Although RSF algorithms are a promising alternative to conventional Cox-PH as data complexity increases, they require a higher number of events for training. In time-to-event analysis, algorithms that consider survival time should be used.

The Effects of Personalizing Colorectal Cancer Risk Communication on Risk Perceptions and Health Behavior Intentions: a Randomized Trial of Average-Risk Adults

Risk assessment tools may help individuals gauge cancer risk and motivate lifestyle and screening behavior changes. Despite the evermore common availability of such tools, little is known about their potential utility in average-risk population approaches to cancer prevention. We evaluated the effects of providing personalized (vs. generic) information concerning colorectal cancer (CRC) risk factors on average-risk individuals' risk perceptions and intentions to engage in three risk-reducing behaviors: CRC screening, diet, and physical activity. Further, we explored whether the receipt of CRC-specific risk assessment feedback influenced individuals' breast cancer risk perceptions and mammography intentions. Using an online survey, N = 419 survey respondents aged 50-75 with no personal or family history of CRC were randomized to receive an average estimate of CRC lifetime risk and risk factor information that was either personalized (treatment) or invariant/non-personalized (control). Respondent risk perceptions and behavioral intentions were ascertained before and after risk assessment administration. No differences were observed in risk perceptions or behavioral intentions by study arm. However, regardless of study arm, CRC screening intentions significantly increased after risk assessment feedback was provided. This occurred despite a significant reduction in risk perceptions. Results support the role simple cancer risk assessment information could play in promoting screening behaviors while improving the accuracy of cancer risk perceptions. Providing cancer risk assessment information may decrease individuals' perceptions of cancer risk to more realistic levels while simultaneously facilitating screening intentions among an average-risk population, regardless of whether provided risk information is personalized.

Latent class regression improves the predictive acuity and clinical utility of survival prognostication amongst chronic heart failure patients

The present study aimed to compare the predictive acuity of latent class regression (LCR) modelling with: standard generalised linear modelling (GLM); and GLMs that include the membership of subgroups/classes (identified through prior latent class analysis; LCA) as alternative or additional candidate predictors. Using real world demographic and clinical data from 1,802 heart failure patients enrolled in the UK-HEART2 cohort, the study found that univariable GLMs using LCA-generated subgroup/class membership as the sole candidate predictor of survival were inferior to standard multivariable GLMs using the same four covariates as those used in the LCA. The inclusion of the LCA subgroup/class membership together with these four covariates as candidate predictors in a multivariable GLM showed no improvement in predictive acuity. In contrast, LCR modelling resulted in a 18-22% improvement in predictive acuity and provided a range of alternative models from which it would be possible to balance predictive acuity against entropy to select models that were optimally suited to improve the efficient allocation of clinical resources to address the differential risk of the outcome (in this instance, survival). These findings provide proof-of-principle that LCR modelling can improve the predictive acuity of GLMs and enhance the clinical utility of their predictions. These improvements warrant further attention and exploration, including the use of alternative techniques (including machine learning algorithms) that are also capable of generating latent class structure while determining outcome predictions, particularly for use with large and routinely collected clinical datasets, and with binary, count and continuous variables.

"Mini Dial-A-Nurses" and "Good Brands": What Are the Desirable Features of Online HIV/STI Risk Calculators?

A wide variety of risk calculators estimate individuals' risk for HIV/sexually transmitted infections (STI) online. These tools can help target HIV/STI screening and optimize clinical decision-making. Yet, little evidence exists on suitable features for these tools to be acceptable to end-users. We investigated the desirable characteristics of risk calculators among STI clinic clients and testing service providers. Participants interacted with online HIV/STI risk calculators featuring varied target audiences, completion lengths, and message outputs. Thematic analysis of focus groups identified six qualities that would make risk calculators more appealing for online client use: providing personalized risk assessments based on users' specific sexual behaviors and HIV/STI-related concerns; incorporating nuanced risk assessment and tailored educational information; supplying quantifiable risk estimates; using non-stigmatizing and inclusive framing; including explanations and next steps; and developing effective and appropriate branding. Incorporating these features in the design of online HIV/STI risk calculators may improve their acceptability among end-users.

Exploring Explanatory Models of Risk in Breast Cancer Risk Counseling Discussions: NSABP/NRG Oncology Decision-Making Project 1

BACKGROUND

Explanatory models represent patient understanding of etiology, pathophysiology, illness, symptoms, and treatments, but little attention has been paid to how they are used by patients "at risk" for future disease.

OBJECTIVE

The aims of this study were to elucidate what constitutes an explanatory model of risk and to describe explanatory models of risk related to developing breast cancer.

METHODS

Thirty qualitative interviews with women identified as at an increased risk for breast cancer were conducted. Interviews were coded to identify domains of explanatory models of risk using a priori codes derived from the explanatory model of illness framework. Within each domain, a grounded thematic analysis described participants' explanatory models related to breast cancer risk.

RESULTS

The domains of treatment and etiology remained similar in a risk context compared with illness, whereas course of illness, symptoms, and pathophysiology differed. We identified a new, integrative concept relative to other domains within explanatory models of risk: social comparisons, which was dominant in risk perhaps due to the lack of physical experiences associated with being "at risk."

CONCLUSIONS

Developing inclusive understandings of risk and its treatment is key to developing a framework for the care of high-risk patients that is both evidence based and sensitive to patient preferences.

IMPLICATIONS FOR PRACTICE

The concept of "social comparisons" can assist healthcare providers in understanding women's decision making under conditions of risk. Ensuring that healthcare providers understand patient perceptions of risk is important because it relates to patient decision making, particularly due to an increasing focus on risk assessment in cancer.

The sociology of cancer: a decade of research

Biomedicine is often presented as the driving force behind improvements in cancer care, with genomics the latest innovation poised to change the meaning, diagnosis, treatment, prevention and lived experience of cancer. Reviewing sociological analyses of a diversity of patient and practitioner experiences and accounts of cancer during the last decade (2007-17), we explore the experiences of, approaches to and understandings of cancer in this period. We identify three key areas of focus: (i) cancer patient experiences and identities; (ii) cancer risk and responsibilities and (iii) bioclinical collectives. We explore these sociological studies of societal and biomedical developments and how sociologists have sought to influence developments in cancer identities, care and research. We end by suggesting that we extend our understanding of innovations in the fields of cancer research to take better account of these wider social and cultural innovations, together with patients, activists' and sociologists' contributions therein.

Risk accuracy of type 2 diabetes in middle aged adults: Associations with sociodemographic, clinical, psychological and behavioural factors

OBJECTIVE

To identify the proportion of individuals with an accurate perception of their risk of type 2 diabetes (T2D) prior to, immediately after and eight weeks after receiving a personalised risk estimate. Additionally, we aimed to explore what factors are associated with underestimation and overestimation immediately post-intervention.

METHODS

Cohort study based on the data collected in the Diabetes Risk Communication Trial. We included 379 participants (mean age 48.9 (SD 7.4) years; 55.1% women) who received a genotypic or phenotypic risk estimate for T2D.

RESULTS

While only 1.3% of participants perceived their risk accurately at baseline, this increased to 24.7% immediately after receiving a risk estimate and then dropped to 7.3% at eight weeks. Those who overestimated their risk at baseline continued to overestimate it, whereas those who underestimated their risk at baseline improved their risk accuracy. We did not identify any other characteristics associated with underestimation or overestimation immediately after receiving a risk estimate.

CONCLUSION

Understanding a received risk estimate is challenging for most participants with many continuing to have inaccurate risk perception after receiving the estimate.

PRACTICE IMPLICATIONS

Individuals who overestimate or underestimate their T2D risk before receiving risk information might require different approaches for altering their risk perception.

My Lived Experiences Are More Important Than Your Probabilities: The Role of Individualized Risk Estimates for Decision Making about Participation in the Study of Tamoxifen and Raloxifene (STAR)

BACKGROUND

Decision-making experts emphasize that understanding and using probabilistic information are important for making informed decisions about medical treatments involving complex risk-benefit tradeoffs. Yet empirical research demonstrates that individuals may not use probabilities when making decisions.

OBJECTIVES

To explore decision making and the use of probabilities for decision making from the perspective of women who were risk-eligible to enroll in the Study of Tamoxifen and Raloxifene (STAR).

METHODS

We conducted narrative interviews with 20 women who agreed to participate in STAR and 20 women who declined. The project was based on a narrative approach. Analysis included the development of summaries of each narrative, and thematic analysis with developing a coding scheme inductively to code all transcripts to identify emerging themes.

RESULTS

Interviewees explained and embedded their STAR decisions within experiences encountered throughout their lives. Such lived experiences included but were not limited to breast cancer family history, a personal history of breast biopsies, and experiences or assumptions about taking tamoxifen or medicines more generally.

CONCLUSIONS

Women's explanations of their decisions about participating in a breast cancer chemoprevention trial were more complex than decision strategies that rely solely on a quantitative risk-benefit analysis of probabilities derived from populations In addition to precise risk information, clinicians and risk communicators should recognize the importance and legitimacy of lived experience in individual decision making.

Desigualdades sociais em saúde e práticas corporais: um exercício singular de análise

Ao reconhecer a relevância dos estudos sobre as desigualdades sociais em saúde, o presente trabalho propõe uma análise que articula essa temática à questão das práticas corporais e sua institucionalização no campo público. Inicia resgatando as perspectivas ético-políticas legitimadoras do ideário da Atenção Primária à saúde e da Promoção da saúde e alguns elementos pontuais do vasto debate teórico sobre as desigualdades e iniquidades sociais em saúde. A partir desses pressupostos problematiza o modo como as práticas corporais têm sido "posicionadas" no universo da Atenção Primária no Brasil, uma vez que sua implementação vem ocorrendo de forma medicalizadora e fragmentada. Para tanto, debruça-se sobre o exame contextual de duas iniciativas públicas existentes no Espírito Santo (ES) que fomentam práticas corporais/atividades físicas. O empreendimento analítico foi subsidiado pelo cotejamento de informações relativas às condições de vida em alguns bairros da cidade de Vitória e às características de tais programas. Finaliza apontando algumas reflexões, com base no quadro empírico-conceitual produzido.

SI RLTD: Risk Scores and Decision Making: The Anatomy of a Decision to Reduce Breast Cancer Risk

AIM: To report the use of a risk score for risk treatment decision-making in women at risk for breast cancer in order to better understand their decision-making situation. BACKGROUND: Tamoxifen and Raloxifene are medications that have been proven to reduce the risk of breast cancer. However, women who understand their personal net benefit from Tamoxifen use chose not to take the medication. To understand this decision, the paper investigates the use of epidemiological risk information in the decision-making process for risk-reducing treatments. METHODS: The narratives of two women are analyzed as they recall their risk score and explain their decision-making process concerning participation in the Study of Tamoxifen and Raloxifene (STAR). Both in-depth interviews follow a narrative approach and were recorded in a U.S. cancer center in 2005. RESULTS: Thinking about risk by analyzing the chances of developing a disease is specific to complex decision-making situations. The associated risk-benefit analysis has to be conducted qualitatively as epidemiological risk information cannot know all details of a woman's life. In addition, a woman's decision is based on the perception of the condition as risk or as disease. Women are willing to treat risk that is perceived as disease, especially when it is based on bodily measurements on which the treatment has an effect. Women are not willing to treat a risk not perceived as disease. CONCLUSION: The net benefit of a treatment as calculated based on epidemiological data cannot easily be translated onto an individual's life. Thus, the complex experience of a woman's life at risk is highly important in decision-making situations. RELEVANCE TO CLINICAL PRACTICE: The ambiguity of statistical risk estimates should be acknowledged and the women's evaluation of her risk valued in risk treatment decision-making.