Economic evaluation for medical artificial intelligence: accuracy vs. cost-effectiveness in a diabetic retinopathy screening case.

Artificial intelligence (AI) models have shown great accuracy in health screening. However, for real-world implementation, high accuracy may not guarantee cost-effectiveness. Improving AI's sensitivity finds more high-risk patients but may raise medical costs while increasing specificity reduces unnecessary referrals but may weaken detection capability. To evaluate the trade-off between AI model performance and the long-running cost-effectiveness, we conducted a cost-effectiveness analysis in a nationwide diabetic retinopathy (DR) screening program in China, comprising 251,535 participants with diabetes over 30 years. We tested a validated AI model in 1100 different diagnostic performances (presented as sensitivity/specificity pairs) and modeled annual screening scenarios. The status quo was defined as the scenario with the most accurate AI performance. The incremental cost-effectiveness ratio (ICER) was calculated for other scenarios against the status quo as cost-effectiveness metrics. Compared to the status quo (sensitivity/specificity: 93.3%/87.7%), six scenarios were cost-saving and seven were cost-effective. To achieve cost-saving or cost-effective, the AI model should reach a minimum sensitivity of 88.2% and specificity of 80.4%. The most cost-effective AI model exhibited higher sensitivity (96.3%) and lower specificity (80.4%) than the status quo. In settings with higher DR prevalence and willingness-to-pay levels, the AI needed higher sensitivity for optimal cost-effectiveness. Urban regions and younger patient groups also required higher sensitivity in AI-based screening. In real-world DR screening, the most accurate AI model may not be the most cost-effective. Cost-effectiveness should be independently evaluated, which is most likely to be affected by the AI's sensitivity.

Time trends and heterogeneity in the disease burden of visual impairment due to cataract, 1990-2019: A global analysis.

Objectives: This study aimed to estimate the disease burden of cataract and evaluate the contributions of risk factors to cataract-associated disability-adjusted life years (DALYs). Materials and methods: Prevalence and DALYs of visual impairment due to cataract were extracted from the Global Burden of Disease (GBD) study 2019 to explore time trends and annual changes. Regional and country-level socioeconomic indexes were obtained from open databases. The time trend of prevalence and DALYs was demonstrated. Stepwise multiple linear regression was used to evaluate associations between the age-standardized rate of DALYs of cataract and potential predictors. Results: Global Prevalence rate of visual impairment due to cataract rose by 58.45% to 1,253.9 per 100,000 population (95% CI: 1,103.3 to 1,417.7 per 100,000 population) in 2019 and the DALYs rate rose by 32.18% from 65.3 per 100,000 population (95% CI: 46.4 to 88.2 per 100,000 population) in 1990 to 86.3 per 100,000 population (95% CI: 61.5 to 116.4 per 100,000 population) in 2019. Stepwise multiple linear regression model showed that higher refractive error prevalence (ß = 0.036, 95% CI: 0.022, 0.050, P < 0.001), lower number of physicians per 10,000 population (ß = -0.959, 95% CI: -1.685, -0.233, P = 0.010), and lower level of HDI (ß = -134.93, 95% CI: -209.84, -60.02, P = 0.001) were associated with a higher disease burden of cataract. Conclusion: Substantial increases in the prevalence of visual impairment and DALYs of cataract were observed from 1990 to 2019. Successful global initiatives targeting improving cataract surgical rate and quality, especially in regions with lower socioeconomic status, is a prerequisite to combating this growing burden of cataract in the aging society.

Burden of near vision loss in China: findings from the Global Burden of Disease Study 2019.

PURPOSE: To investigate the burden of near vision loss (NVL) in China by year, age and gender from 1990 to 2019. METHODS: We used estimates from the Global Burden of Disease (GBD) 2019 study to report the prevalence and disability-adjusted life-years (DALYs) due to NVL in China. Estimates of crude counts and age-standardised rates per 100 000 population are accompanied by 95% uncertainty intervals (UIs). We summarised the age-specific and sex-specific patterns and trends regarding the burden of NVL in China, compared with seven neighbouring countries. RESULTS: From 1990 to 2019, the all-age number and rate for NVL prevalence and DALYs increased significantly in China (all p<0.001). The age-standardised rate decreased from 7538.1 (95% UI 6946.3 to 8075.1) to 7392.9 (95% UI 6855.8 to 7890.5) per 100 000 population for NVL prevalence (p=0.107), and from 74.9 (95% UI 69.6 to 79.9) to 73.8 (95% UI 70.6 to 80.1) per 100 000 population for DALYs (p=0.388). Women had higher NVL prevalence (t=170.1, p<0.001) and DALYs (t=192.5, p<0.001) than men. Higher disease burden of NVL was observed in the middle-aged and elderly population. The age-standardised prevalence and DALY rate attributable to NVL in China were lower than in India, North Korea, Pakistan (all p<0.001), but higher than Russia, South Korea, Singapore and Japan (all p<0.001). CONCLUSIONS: Despite a small decrease in age-standardised prevalence and DALYs due to NVL in China in the past two decades, the existing burden is still considerable and significantly higher compared with neighbouring developed countries. An approach that includes all stakeholders is needed to further reduce this burden.

Enrollee Experience with Providers in the Arkansas Medicaid Expansion Program.

BACKGROUND: Patient ratings of their healthcare experience as a quality measure have become critically important since the implementation of the Affordable Care Act (ACA). The ACA enabled states to expand Medicaid eligibility to reduce uninsurance nationally. Arkansas gained approval to use Medicaid funds to purchase a qualified health plan (QHP) through the ACA marketplace for newly eligible beneficiaries. OBJECTIVE: We compare patient-reported satisfaction between fee-for-service Medicaid and QHP participants. DESIGN: The Consumer Assessment of Healthcare Providers and Systems (CAHPS) was used to identify differences in Medicaid and QHP enrollee healthcare experiences. Data were analyzed using a regression discontinuity design. PARTICIPANTS: Newly eligible Medicaid expansion participants enrolled in Medicaid during 2013 completed the Consumer Assessment of Health Providers and Systems (CAHPS) survey in 2014. Survey data was analyzed for 3156 participants (n = 1759 QHP/1397 Medicaid). MEASURES: Measures included rating of personal and specialist provider, rating of all healthcare received, and whether the provider offered to communicate electronically. Demographic and clinical characteristics of the enrollees were controlled for in the analyses. METHODS: Regression-discontinuity analysis was used to evaluate differential program effects on positive ratings as measured by the CAHPS survey while controlling for demographic and health characteristics of participants. KEY RESULTS: Adjusted logistic regression models for overall healthcare (OR = 0.71, 95%CI = 0.56-0.90, p = 0.004) and personal doctor (OR = 0.68, 95%CI = 0.53-0.87, p = 0.002) predicted greater satisfaction among QHP versus Medicaid participants. Results were not significant for specialists or for use of electronic communication with provider. CONCLUSIONS: Using a quasi-experimental statistical approach, we were able to control for observed and unobserved heterogeneity showing that among participants with similar characteristics, including income, QHP participants rated their personal providers and healthcare higher than those enrolled in Medicaid. Access to care, utilization of care, and healthcare and health insurance literacy may be contributing factors to these results.

Higher Rates of Preventive Health Care With Commercial Insurance Compared With Medicaid: Findings From the Arkansas Health Care Independence "Private Option" Program.

BACKGROUND: A requirement of the Arkansas Medicaid Section 1115 demonstration waiver was to evaluate the level of care received for Medicaid expansion eligible beneficiaries enrolled in commercial Qualified Health Plans (QHPs) in the Health Care Independence "Private Option" Program. This allowed for a direct comparison of Medicaid and commercial system performance serving similar newly covered adults. RESEARCH DESIGN: In 2014, assignment to either Medicaid or a QHP was made based upon a psychometrically derived continuous composite score to exceptional health care needs assessment screener using a sharp a priori threshold cutpoint. Using a regression discontinuity design we compared preventive care (flu vaccination and screening rates) services in the 2 programs over 3 years. RESULTS: Compared with Medicaid enrollees, a higher percentage of QHP enrollees consistently received eligible preventive care screenings with 15.3, and 6.9% more receiving at least 1 or all eligible screenings, respectively. For individual preventive care outcomes and compared with Medicaid enrollees over the 3 years under study, a higher percentage of eligible QHP enrollees received a flu shot, cholesterol screenings, glycated hemoglobin assessment, and cervical and breast cancer periodic assessments. No differences were found for colorectal periodic assessments. CONCLUSIONS: These findings suggest that at least for preventive services, the Medicaid federal equal access requirement is not being met for those within Medicaid fee-for-service coverage. This persisted across all 3 years of the program. Differential payment rates for services between Medicaid and QHPs are likely a major contributing factor.

Utilisation and perceptions towards smart device visual acuity assessment in Australia: a mixed methods approach.

OBJECTIVES: To investigate mobile health product use in Australia and societal and clinician perceptions towards smartphone based visual acuity (VA) assessment tools. DESIGN: Quantitative analysis of a cross-sectional survey delivered to the general public and thematic analysis of in-depth interviews of eye health clinicians. SETTING: Online survey within Australia and face-to-face in-depth interviews of clinicians. PARTICIPANTS: 1016 adults were recruited via Survey Monkey Audience, social media (Facebook and Twitter), Rotary Australia and Lions Clubs Australia. Six clinicians were recruited from private and public settings in Melbourne, Australia. PRIMARY AND SECONDARY OUTCOME MEASURES: The study assessed socio-demographic characteristics, history of mobile health product use and perceived advantages and potential drawbacks of smartphone based VA assessment tools. RESULTS: A total of 14.4% of the study population had previously used a mobile-based health product. After adjusting for covariates, younger age (p=0.001), male gender (p=0.01) and higher income (>$45 000) were associated with increased likelihood of having used a mobile health product (p=0.005). Seventy-two per cent of participants would use an automated smartphone based VA assessment tool, provided that the accuracy was on par to that of human assessors. Convenience (37.3%) and cost-savings (15.5%) were ranked as the greatest perceived advantages. While test accuracy (50.6%), a lack of personal contact with healthcare providers (18.3%) and data security (11.9%) were the greatest concerns. Themes to emerge from clinician qualitative data included the potential benefits for identifying refractive error in patients, as well as the ability to self-monitor vision. Concerns were raised over the potential misuse of self-testing vision apps and the inability to detect pathology. CONCLUSION: Our findings suggest that a substantial proportion of the Australian population do not use mobile health products. Furthermore, there remains notable concerns, including test accuracy and data privacy, with smartphone-based VA assessment tools by both clinicians and the public.

Does a quality improvement intervention for anxiety result in differential outcomes for lower-income patients?

OBJECTIVE: The authors examined the effects of a collaborative care intervention for anxiety disorders in primary care on lower-income participants relative to those with higher incomes. They hypothesized that lower-income individuals would show less improvement or improve at a lower rate, given that they would experience greater economic stress over the treatment course. An alternative hypothesis was that lower-income participants would improve at a higher rate because the intervention facilitates access to evidence-based treatment, which typically is less available to persons with lower incomes. METHOD: Baseline demographic and clinical characteristics of patients with lower (N=287) and higher (N=717) income were compared using t tests and chi-square tests for continuous and categorical variables, respectively. For the longitudinal analysis of intervention effects by income group, the outcome measures were jointly modeled at baseline and at 6, 12, and 18 months by study site, income, time, intervention, time and intervention, income and time, income and intervention, and time, intervention, and income. RESULTS: Although lower-income participants were more ill and had greater disability at baseline than those with higher incomes, the two income groups were similar in clinical response. The lower-income participants experienced a comparable degree of clinical improvement, despite receiving fewer treatment sessions, less relapse prevention, and less continuous care. CONCLUSIONS: These findings contribute to the ongoing discussion as to whether or not, and to what extent, quality improvement interventions work equally well across income groups or require tailoring for specific vulnerable populations.

Religiosity and decreased risk of substance use disorders: is the effect mediated by social support or mental health status?

OBJECTIVE: The negative association between religiosity (religious beliefs and church attendance) and the likelihood of substance use disorders is well established, but the mechanism(s) remain poorly understood. We investigated whether this association was mediated by social support or mental health status. METHOD: We utilized cross-sectional data from the 2002 National Survey on Drug Use and Health (n = 36,370). We first used logistic regression to regress any alcohol use in the past year on sociodemographic and religiosity variables. Then, among individuals who drank in the past year, we regressed past year alcohol abuse/dependence on sociodemographic and religiosity variables. To investigate whether social support mediated the association between religiosity and alcohol use and alcohol abuse/dependence we repeated the above models, adding the social support variables. To the extent that these added predictors modified the magnitude of the effect of the religiosity variables, we interpreted social support as a possible mediator. We also formally tested for mediation using path analysis. We investigated the possible mediating role of mental health status analogously. Parallel sets of analyses were conducted for any drug use, and drug abuse/dependence among those using any drugs as the dependent variables. RESULTS: The addition of social support and mental health status variables to logistic regression models had little effect on the magnitude of the religiosity coefficients in any of the models. While some of the tests of mediation were significant in the path analyses, the results were not always in the expected direction, and the magnitude of the effects was small. CONCLUSIONS: The association between religiosity and decreased likelihood of a substance use disorder does not appear to be substantively mediated by either social support or mental health status.

Patterns of health services use associated with Veterans Affairs outpatient substance-use treatment.

OBJECTIVE: This study examined service use and costs associated with receiving intensive outpatient (IOP) substance-use treatment. METHOD: Veterans Affairs national databases were used to identify all veterans receiving IOP substance-use treatment in fiscal year (FY) 1999. Changes in the use of health services (2 years pretreatment compared with 2 years posttreatment) were examined. Three IOP treatment groups were defined: (1) veterans attending 1-5 IOP visits (n=2,384), (2) 6-14 IOP visits (n=2,940), and (3) >or= 15 IOP visits (n=3,005). In addition, a demographically matched no-treatment group was defined as veterans in primary care with a substance-use diagnosis and no substance-use treatment in FY 1999 (n = 7,328). General medical, psychiatric, and substance-use encounters in inpatient and outpatient settings were retrieved for the 4 years. Generalized estimating equations analyses were used to compare changes in service use across the three IOP groups controlling for demographic characteristics and comorbidity. RESULTS: Total outpatient visits increased in all three treatment groups after IOP, with the largest increase observed in the group with >or= 15 visits (mean visits = 5, 44, and 57, respectively). In comparison, total visits decreased in the no-treatment group (-7). This pattern was similar across settings (general medical, psychiatric, and substance use). Total inpatient days decreased among the four groups, with the no-treatment group having the least amount of decrease (mean change in days = .34, .32, .18, and .03, respectively). Analysis of cost data showed similar findings. CONCLUSIONS: Results indicated that more intense IOP treatment is associated with increases in subsequent psychiatric, substance-use, medical, and total outpatient care and decreases in expensive inpatient care.

Measuring use of health services for at-risk drinkers: how brief can you get?

This study examines the validity, utility, and costs of using a brief telephone-administered instrument, the Brief Health Services Questionnaire (BHSQ), for self-reported health care provider contacts relative to collection and abstraction of complete medical records. The study sample was 441 community-dwelling at-risk drinkers who participated in an 18-month longitudinal study. Agreement between BHSQ self-reports and abstracted provider contacts was good to very good for general medical (79% agreement, kappa = .50) and specialty mental health contacts (93% agreement, kappa = .62), but low for "other" miscellaneous health contacts (61% agreement, kappa = .04). Average cost to collect and abstract complete medical records was 424 US dollars per study participant, whereas average cost to administer only the BHSQ was 31 US dollars per participant. Although it is not possible to conduct a formal cost-effectiveness analysis, results suggest the BHSQ is a viable option for collecting self-reported health provider contacts in a sample of at-risk drinkers, with definite cost advantages over more elaborate data collection methods.

Health services utilization and cost for at-risk drinkers: rural and urban comparisons.

OBJECTIVE: The purpose of this study was to examine differences between healthcare use and associated costs in rural and urban at-risk drinkers. METHOD: Primary healthcare utilization and cost data were collected from 1995 to 1998 on rural (n = 215) and urban (n = 228) cohorts of drinkers residing in six southeastern states who met criteria for at-risk drinking. Data were obtained through subject interview and from abstracts of medical and pharmacy records. RESULTS: Overall healthcare costs were not significantly different between the rural and urban cohorts. For subjects who incurred any hospital costs (including emergency room [ER] visits), however, costs were significantly greater (p < .01) for rural patients (median = dollars 2,561) than for urban patients (median = dollars 865). Hospital costs associated with patients' ER visits and any subsequent admissions were also greater (p < .01) for rural patients (median = dollars 1,004) than for urban patients (median = dollars 512). Use of healthcare services was significantly more likely to occur among women (p < .0001), individuals with lower overall self-reported physical health (p < .01) and individuals with health insurance (p < .0001). Among subjects who used healthcare services, greater costs were significantly associated with older age (p < .05), being female (p <.0001), having lower overall physical health (p < .0001) and having health insurance (p < .01). CONCLUSIONS: While overall healthcare costs are not significantly different between rural and urban residents in this sample of at-risk drinkers, there are some notable differences in the costs associated with inpatient and ER services.

Rural-urban differences in health care benefits of a community-based sample of at-risk drinkers.

CONTEXT: Different types of health plan cost-containment strategies (eg, gatekeeping, selective contracting, and cost-sharing) may affect the utilization of behavioral health services differently in urban and rural areas. PURPOSE: This research compares the cost-containment strategies used by the health plans of insured at-risk drinkers residing in rural and urban areas. METHODS: A screening instrument for at-risk drinking was administered by phone to approximately 12,000 residents of 6 southern states; 442 at-risk drinkers completed 4 interviews over a 2-year period and consented to release insurance and medical records. Two thirds of the sample (n=294) were insured during the last 6 months of the study. In 1998, health plan characteristics were successfully collected for 217 (72.3%) of the insured at-risk drinkers, representing 113 different health plans and 206 different policies. FINDINGS: Compared with urban at-risk drinkers, rural at-risk drinkers were significantly less likely to be enrolled in a health plan with gatekeeping policies for both behavioral health (P = .001), and physical health (P = .031). Compared with urban enrollees, rural enrollees were significantly more likely to pay deductibles (P = .042), to pay coinsurance for physical health services (P = .002), and to have limits placed on physical health services use (P = .067), but they were less likely to pay copayments for physical health (P = .046). Rural enrollees were less likely to face higher copayments (P = .007) and higher coinsurance (P = .076) for mental health than for physical health, compared to urban enrollees. CONCLUSIONS: Because rural residents were more likely to be enrolled in indemnity plans and less likely to be enrolled in health maintenance organizations, rural at-risk drinkers were enrolled in plans that relied less on supply-side cost-containment strategies and more on demand-side cost-containment strategies targeting physical health service use, compared with their urban counterparts. Rural at-risk drinkers were less likely to be enrolled in health plans with greater cost-sharing for mental health than for physical health compared to urban at-risk drinkers.

Differences between physical and behavioral health benefits in the health plans of at-risk drinkers.

OBJECTIVE: The goal of this study was to describe the physical and behavioral health benefits of a representative community-based sample of at-risk drinkers potentially in need of behavioral health services. METHODS: A screening instrument for at-risk drinking was administered by telephone to a random community sample of more than 12,000 adults. A telephone interview was conducted with the health plans of 294 at-risk drinkers who were insured and who consented to the release of their insurance records to collect information about supply-side cost-containment strategies (for example, gatekeeping and restrictions on choice of provider), and demand-side cost-containment strategies (for example, deductibles, limits, coinsurance, and copayments). Information about health plan characteristics was successfully collected for 217 (72 percent) of the insured at-risk drinkers, representing 113 different health plans and 206 different policies. RESULTS: Both provider choice restrictions and gatekeeping were more likely to be used for behavioral health care than for physical health care. Greater cost-sharing for mental health than for physical health was most often achieved by using additional limits (83 percent) and higher coinsurance (66 percent) and less often achieved by using higher copayments (38 percent) and additional deductibles (13 percent). The greater cost-sharing for behavioral health amounted to a 30 percent ($42) difference in annual out-of-pocket costs for an average user of behavioral health services compared with full parity. CONCLUSIONS: The results provide information to advocacy groups and policy makers about how much equalization would have to occur in the insurance market before full parity could be achieved between physical health and behavioral health benefits for a population of individuals potentially in need of behavioral health services.