Understanding variation in chronic disease outcomes

Spoiled for Choice? Personalized Recommendation for Healthcare Decisions: A Multiarmed Bandit Approach

Choice overload is a common problem in many online settings, including healthcare. Online healthcare platforms tend to provide a large variety of behavior intervention information or programs to help individuals modify their lifestyles to improve wellness. However, having too many options can significantly increase searching cost, prevent users from discovering the truly relevant interventions, and harm users’ long-term healthcare decision-making efficiency. This motivates us to propose a personalized healthcare recommendation system to provide tailored support for individuals’ intervention participation. The proposed framework, a deep-learning and diversity-enhanced multiarmed bandit (DLDE-MAB), integrates several predictive and prescriptive analytics components to combat the unique challenges presented in the healthcare recommendation setting. It leverages online machine learning to provide adaptive and real-time support, a theory-guided diversity promotion scheme to cover multiple healthcare needs, and deep learning to further enhance dynamic context representation. Through extensive experiments, we show that the proposed framework outperforms various competing models in terms of its adaptivity to data dynamics, diversity, and uncertainty. The proposed model and evaluation results provide important implications for business intelligence and personalized, contextualized, and agile healthcare decision making.

Análisis evolutivo del concepto adaptación a la diabetes tipo 2

El concepto de adaptación tiene diferentes aplicaciones y acepciones. En personas que viven con diabetes tipo 2 (DT2) se asocia con la observancia del tratamiento y el control glucémico. Es utilizado frecuentemente en la literatura de salud, pero no está claro qué significa en personas que viven con esta enfermedad, por lo que el propósito de esta investigación fue analizar el concepto de adaptación a la DT2. Metodología: se utilizó el método de análisis evolutivo de concepto de Rodgers. Los datos se codificaron por autor, 16 en total, palabras clave, hallazgos, antecedentes, atributos y consecuencias, conceptos relativos y sustitutos. La guía del análisis respondió a tres preguntas: ¿cómo define el concepto el autor? ¿Qué características o atributos se establecen? ¿Qué idea da el autor acerca del concepto? Resultados: el concepto de adaptación muestra atributos que permiten definirlo como un proceso que se presenta en tres fases: reacción, asimilación y respuesta. Los antecedentes corresponden a conductas previas al proceso, los resultados incluyen desafíos y cambios. Conclusiones: los atributos, los antecedentes y las consecuencias ofrecen guías para la investigación; es necesario ampliar el estudio en el contexto del hogar, la familia, el trabajo y el hospital.

Personalised physician learning intervention to improve hypertension and lipid control: randomised trial comparing two methods of physician profiling

PURPOSE

To assess the impact of personalised physician learning (PPL) interventions using simulated learning cases on control of hypertension and dyslipidaemia in primary care settings.

METHODS

A total of 132 primary care physicians, 4568 eligible patients with uncontrolled hypertension, and 15 392 eligible patients with uncontrolled dyslipidaemia were cluster-randomised to one of three conditions: (a) no intervention, (b) PPL-electronic medical record (EMR) intervention in which 12 PPL cases were assigned to each physician based on observed patterns of care in the EMR in the previous year, or (c) PPL-ASSESS intervention in which 12 PPL cases were assigned to each physician based on their performance on four standardised assessment cases. General and generalised linear mixed models were used to account for clustering and to model differences in patient outcomes in the study arms.

RESULTS

Among patients with uncontrolled hypertension at baseline, 49.1%, 46.6% and 47.3% (p=0.43) achieved blood pressure (BP) targets at follow-up. Among patients with uncontrolled dyslipidaemia at baseline, 37.5%, 37.3% and 38.1% (p=0.72) achieved low density lipoprotein cholesterol targets at follow-up in PPL-EMR, PPL-ASSESS and the control group, respectively. Although systolic (BP) (p<0.001) and lipid (p<0.001) values significantly improved during the study, the group-by-time interaction term showed no differential change in systolic BP values (p=0.51) or lipid values (p=0.61) among the three study arms. No difference in intervention effect was noted when comparing the PPL-EMR with the PPL-ASSESS intervention (p=0.47).

CONCLUSIONS

The two PPL interventions tested in this study did not lead to improved control of hypertension or dyslipidaemia in primary care clinics during a mean 14-month follow-up period. This null result may have been due in part to substantial overall improvement in BP and lipid control at the study sites during the study.

TRIAL REGISTRATION NUMBER

NCT00903071.

Improving the quality of rheumatoid arthritis patients' education using written information

OBJECTIVE

The objective of this study was to evaluate whether the quality of patient education could be improved by using written education materials.

METHODS

Seventy-five inpatients with rheumatoid arthritis (RA) were provided with individual education sessions during their inpatient stay. The education sessions were supported with written educational materials. A patient education quality instrument was used to assess the sufficiency and implementation of the education and the readiness of nurses and doctors to deliver the education. A Mann-Whitney U-test and content analysis was used to analyse the data.

RESULTS

There was a statistically significant difference in the sufficiency of education concerning the disease, medication and treatment after the revised material was introduced (p < 0.005). The quality of interaction improved significantly during the intervention (p = 0.004). The strengths of the education included individual treatment, two-way interaction, the opportunity to receive patient education and its sufficiency.

CONCLUSION

Clear, readable and understandable written education material improved the quality of the education of RA patients in terms of implementation, sufficiency and the readiness of nurses and doctors to deliver the education.

Is there a downside to customizing care? Implications of general and patient-specific treatment strategies

RATIONALE, AIMS AND OBJECTIVES

The use of general clinical guidelines versus customization of patient care presents a dilemma for clinicians managing chronic illness. The objective of this project is to investigate the claim that the performance of customized strategies for the management of chronic illness depends on accurate patient categorization, and inaccurate categorization can lead to worse performance than that achievable using a general clinical guideline.

METHODS

This paper is based on an analysis of a basic utility model that differentiates between the use of general management strategies and customized strategies.

RESULTS

The analysis identifies necessary conditions for preferring general strategies to customized strategies as a trade-off between strategy performance and the probability of correct patient categorization. The analysis shows that customized treatment strategies developed under optimal conditions are not necessarily preferred.

CONCLUSIONS

Results of the analysis have four implications regarding the design and use of clinical guidelines and customization of care: (i) the balance between the applications of more general strategies versus customization depends on the specificity and accuracy of the strategies; (ii) adoption of clinical guidelines may be stifled as the complexity of guidelines increases to account for growing evidence; (iii) clinical inertia (i.e. the failure to intensify an indicated treatment) can be a rational response to strategy specificity and the probability of misapplication; and, (iv) current clinical guidelines and other decision-support tools may be improved if they accommodate the need for customization of strategies for some patients while providing support for proper categorization of patients.

Simulated physician learning intervention to improve safety and quality of diabetes care: a randomized trial

OBJECTIVE

To assess two physician learning interventions designed to improve safety and quality of diabetes care delivered by primary care physicians (PCPs).

RESEARCH DESIGN AND METHODS

This group randomized clinical trial included 57 consenting PCPs and their 2,020 eligible adult patients with diabetes. Physicians were randomized to no intervention (group A), a simulated case-based physician learning intervention (group B), or the same simulated case-based learning intervention with physician opinion leader feedback (group C). Dependent variables included A1C values, LDL cholesterol values, pharmacotherapy intensification rates in patients not at clinical goals, and risky prescribing events.

RESULTS

Groups B and C had substantial reductions in risky prescribing of metformin in patients with renal impairment (P = 0.03). Compared with groups A and C, physicians in group B achieved slightly better glycemic control (P = 0.04), but physician intensification of oral glucose-lowering medications was not affected by interventions (P = 0.41). Lipid management improved over time (P < 0.001) but did not differ across study groups (P = 0.67).

CONCLUSIONS

A simulated, case-based learning intervention for physicians significantly reduced risky prescribing events and marginally improved glycemic control in actual patients. The addition of opinion leader feedback did not improve the learning intervention. Refinement and further development of this approach is warranted.

Randomized trial of quality improvement intervention to improve diabetes care in primary care settings

OBJECTIVE

To assess the impact of a quality improvement (QI) intervention on the quality of diabetes care at primary care clinics.

RESEARCH DESIGN AND METHODS

Twelve primary care medical practices were matched by size and location and randomized to intervention or control conditions. Intervention clinic staff were trained in a seven-step QI change process to improve diabetes care. Surveys and medical record reviews of 754 patients, surveys of 329 clinic staff, interviews with clinic leaders, and analysis of training session videotapes evaluated compliance with and impact of the intervention. Mixed-model nested analyses compared differences in the quality of diabetes care before and after intervention.

RESULTS

All intervention clinics completed at least six steps of the seven-step QI change process in an 18-month period and, compared with control clinics, had broader staff participation in QI activities (P = 0.04), used patient registries more often (P = 0.03), and had better test rates for HbA(1c) (A1C), LDL, and blood pressure (P = 0.02). Other processes of diabetes care were unchanged. The intervention did not improve A1C (P = 0.54), LDL (P = 0.46), or blood pressure (P = 0.69) levels or a composite of these outcomes (P = 0.35).

CONCLUSIONS

This QI change process was successfully implemented but failed to improve A1C, LDL, or blood pressure levels. Data suggest that to be successful, such a QI change process should direct more attention to specific clinical actions, such as drug intensification and patient activation.

Characteristics of adults who use prayer as an alternative therapy

PURPOSE

To describe the demographics, health-related and preventive-health behaviors, health status, and health care charges of adults who do and do not pray for health.

DESIGN

Cross-sectional survey with 1-year follow-up.

SETTING

A Minnesota health plan.

SUBJECTS

A stratified random sample of 5107 members age 40 and over with analysis based on 4404 survey respondents (86%).

MEASURES

Survey data included health risks, health practices, use of preventive health services, satisfaction with care, and use of alternative therapies. Health care charges were obtained from administrative data.

RESULTS

Overall, 47.2% of study subjects reported that they pray for health, and 90.3% of these believed prayer improved their health. After adjustment for demographics, those who pray had significantly less smoking and alcohol use and more preventive care visits, influenza immunizations, vegetable intake, satisfaction with care, and social support and were more likely to have a regular primary care provider. Rates of functional impairment, depressive symptoms, chronic diseases, and total health care charges were not related to prayer

CONCLUSIONS

Those who pray had more favorable health-related behaviors, preventive service use, and satisfaction with care. Discussion of prayer could help guide customization of clinical care. Research that examines the effect of prayer on health status should adjust for variables related both to use of prayer and to health status.

Predictors of health care costs in adults with diabetes

OBJECTIVE

The purpose of this study was to assess the impact of baseline A1c, cardiovascular disease, and depression on subsequent health care costs among adults with diabetes.

RESEARCH DESIGN AND METHODS

A prospective analysis was performed of data from a patient survey and medical record review merged with 3 years of medical claims. Costs were estimated using detailed data on resource use and Medicare payment methodologies. Generalized linear models were used to analyze costs related to clinical predictors after adjusting for demographic and socioeconomic factors.

RESULTS

In multivariate analysis of 1,694 adults with diabetes, 3-year costs in those with coronary heart disease (CHD) and hypertension were over 300% of those with diabetes only (46,879 dollars vs. 14,233 dollars; P < 0.05). Depression was associated with a 50% increase in costs (31,967 dollars vs. 21,609 dollars; P < 0.05). Relative to those with a baseline A1c of 6%, those with an A1c of 10% had 3-year costs that were 11% higher (26,408 dollars vs. 23,873 dollars; P < 0.05). Higher A1c predicted higher costs only for those with baseline A1c >7.5% (P = 0.015).

CONCLUSIONS

In adults with diabetes, CHD, hypertension, and depression spectrum disorders more strongly predicted future costs than the A1c level. Concurrent with aggressive efforts to control glucose, greater efforts to prevent or control CHD, hypertension, and depression are necessary to control health care costs in adults with diabetes.

Making improvements in the management of patients with type 2 diabetes: a possible role for the control of variation in glycated hemoglobin

Glucose level varies over time due to a number of complex physiologic processes. Evidence suggests variation in glucose level contributes to risk of complications. The timescale associated with variation in glucose level is on the order of seconds to minutes, yet diabetes complications stem from years of cumulative effects. This difference between timescale suggests a slower timescale may better represent the influential component of variation. We hypothesize variation in glycated hemoglobin captures the component of variation associated with future complications. Moreover, we hypothesize that patient-management strategies influence variation in glycated hemoglobin level. From a systems control perspective, increasing variation may well reflect a policy of closed loop feedback control where changes in patient glycated hemoglobin are addressed after the fact. Such a strategy attends to problems as they arise. In contrast, decreasing variation may result from a clinical strategy that is anticipatory and proactive. A physician using a proactive strategy will base current moves on anticipation of future states, controlling variation in patient outcomes such as glycated hemoglobin. We motivate our discussion using observational data from a large multispecialty medical group in Minnesota: we characterize the within-patient trend and variation of glycated hemoglobin in adults with type 2 diabetes, describe patterns of variation, and identify factors associated with variation. Our hypotheses imply: (1) patterns of variation in glycated hemoglobin reflect physician treatment strategy; (2) variation provides an independent contribution to risk of diabetes complications; (3) the development of treatment strategies that control variation may be a beneficial goal in the management of type 2 diabetes.

A structured review of the effect of economic incentives on consumers' preventive behavior

Improving participation in preventive activities will require finding methods to encourage consumers to engage in and remain in such efforts. This review assesses the effects of economic incentives on consumers' preventive health behaviors. A study was classified as complex preventive health if a sustained behavior change was required of the consumer; if it could be accomplished directly (e.g., immunizations), it was considered simple. A systematic literature review identified 111 randomized controlled trials of which 47 (published between 1966 and 2002) met the criteria for review. The economic incentives worked 73% of the time (74% for simple, and 72% for complex). Rates varied by the goal of the incentive. Incentives that increased ability to purchase the preventive service worked better than more diffuse incentives, but the type matters less than the nature of the incentive. Economic incentives are effective in the short run for simple preventive care, and distinct, well-defined behavioral goals. Small incentives can produce finite changes, but it is not clear what size of incentive is needed to yield a major sustained effect.

Is patient readiness to change a predictor of improved glycemic control?

OBJECTIVE

To test the hypothesis that patient readiness to change (RTC) predicts future changes in glycemic control in adults with diabetes.

RESEARCH DESIGN AND METHODS

We linked survey data with HbA1c data for a stratified random sample of consenting adults with diabetes. Change in HbA1c from baseline to the 1-year follow-up was computed and used as a dependent variable. Linear regression models assessed RTC and other patient variables as predictors of HbA1c change.

RESULTS

Among 617 patients with baseline HbA1c > or = 7% and complete data for analysis, RTC predicted subsequent improvement in HbA1c for those with higher physical functioning (interaction t = -2.45, P < 0.05). Other factors that predicted HbA1c improvement in multivariate linear regression models included higher self-reported medication adherence (t = -4.41, P < 0.01), higher baseline HbA1c (t = -15.08, P < 0.01), and older age (t = -2.61, P < 0.01).

CONCLUSIONS

Diabetes RTC independently predicts change in HbA1c for patients with high but not for patients with low functional health status. Customized use of RTC assessment may have potential to improve care.

Key issues in transforming health care organizations for quality: the case of advanced access

BACKGROUND

The 2001 Institute of Medicine (IOM) report highlighted the need for transformation of the U.S. health care system. This rigorous qualitative evaluation of transformational change for patient access in one large multispecialty group practice identifies the major issues facing organizations addressing the IOM challenge.

METHODS

Semistructured depth interviews were conducted with the medical and administrative leaders at all levels, physicians, and nurses from 17 primary care clinics in one integrated medical group two years after they began to transform their approach to primary care patient appointment access.

RESULTS

The mean time to third-next-available appointment was reduced by 76% during one year, from 17.8 days to 4.2 days. Nine important issues related to the change process were identified from clinic interviews. When combined with issues identified by central leaders, 13 themes stood out as lessons in transformational change. A major issue is the tension between physician autonomy and both effective organizational function and putting patients first. Physician autonomy is also diminished by the need to standardize and systematize care.

CONCLUSIONS

Transformational change in care delivery is possible in large and complex group practices. Changes that directly affect care delivery and physician autonomy present particular challenges to physicians that need to be attended to if the changes are to be successful.

Variation in diabetes care by age: opportunities for customization of care

Background

The quality of diabetes care provided to older adults has usually been judged to be poor, but few data provide direct comparison to other age groups. In this study, we hypothesized that adults age 65 and over receive lower quality diabetes care than adults age 45–64 years old.

Methods

We conducted a cohort study of members of a health plan cared for by multiple medical groups in Minnesota. Study subjects were a random sample of 1109 adults age 45 and over with an established diagnosis of diabetes using a diabetes identification method with estimated sensitivity 0.91 and positive predictive value 0.94. Survey data (response rate 86.2%) and administrative databases were used to assess diabetes severity, glycemic control, quality of life, microvascular and macrovascular risks and complications, preventive care, utilization, and perceptions of diabetes.

Results

Compared to those aged 45–64 years (N = 627), those 65 and older (N = 482) had better glycemic control, better health-related behaviors, and perceived less adverse impacts of diabetes on their quality of life despite longer duration of diabetes and a prevalence of cardiovascular disease twice that of younger patients. Older patients did not ascribe heart disease to their diabetes. Younger adults often had explanatory models of diabetes that interfere with effective and aggressive care, and accessed care less frequently. Overall, only 37% of patients were simultaneously up-to-date on eye exams, foot exams, and glycated hemoglobin (A1c) tests within one year.

Conclusion

These data demonstrate the need for further improvement in diabetes care for all patients, and suggest that customisation of care based on age and explanatory models of diabetes may be an improvement strategy that merits further evaluation.