Question: Do OurHealth primary care clinics improve health & reduce healthcare costs? OurHealth Patient Engagement Analysis June 2018 An examination of health improvements, utilization & cost of care for employer-based primary care clinics ourhealth.org 866-434-3255
--- Executive Summary Do patients who engage with OurHealth employer-based primary care services improve their health and reduce healthcare costs? To answer this question, we analyzed three employers of varied size and industry whose medical plan members engaged with OurHealth s onsite and near-site primary care clinic services. Patients who engaged with OurHealth experienced greater health improvements than those who didn t, even when adjusted for baseline health status. These engaged patients also created greater savings in the form of reduced healthcare claims costs, including $2.6 million in client savings in two years. Other key findings include: Health improvements: 47.4% of identified prediabetics (1,448 patients) experienced health improvements that moved their hemoglobin A1c (HbA1c) * values into optimal range, resulting in a projected savings of $1.7 million. Additionally, 12.1% of diabetics identified (555 patients) with high HbA1c values (between 6.5 and 9.0) and 9.9% of diabetics identified (182 patients) with very high HbA1c values (9.0 or above) experienced health improvements that moved their HbA1c values into optimal range. Identifying Prediabetes Financial Impact Prediabetics Identified 3,052 % Progressing to Type 2 15% Estimated Annual Cost of Type 2 Diabetes Patient $7,900 Total Potential Cost of Prediabetics advancing to Type 2 $ 3,616,620 % of Prediabetics improved to optimal range 47.4% Total Estimated Savings $1,715,880 Savings per Prediabetic Identified $562.21 * HbA1c correlates to average blood glucose, or blood sugar, values over the previous 90 days. CDC estimate of prediabetics that will advance to Type 2 diabetes Average cost to cover an employee with diabetes per Johns Hopkins Medicine 2
Patients who engaged in health coaching reduced their HbA1c values at a rate 67.3% greater than patients who did not engage with health coaching. Additionally, more numerous health coaching and provider visits correlated to more significant HbA1c reductions. Patients experienced a cumulative reduction of more than 18,200 low-density lipoprotein (LDL cholesterol) * points. Patients experienced a cumulative reduction of more than 8,800 total cholesterol points. Emergency Department utilization: Patients who engaged with OurHealth used the Emergency Department 22.8% less than patients who did not engage (for all causes). At an average ER visit cost of $1,500, this reduction resulted in a projected savings of $652,000. Cost of care: When patients engaged with OurHealth, their observed medical and drug claims costs were 17% lower than expected. Over two years, this trend resulted in a total client savings of more than $2.6 million or $420 per member per year. *LDL cholesterol is known as the "bad" cholesterol and can indicate an increased risk for heart attack, stroke, and other conditions. Consumer Health Ratings, Emergency Room Typical Average Cost of Hospital ED Visit To learn more about OurHealth s data analysis and improved outcomes for patients and employers, visit ourhealth.org 3
--- Background In the United States in 2016, more than $3.3 trillion was spent on healthcare. By 2026 nearly 1 out of every 5 dollars, or $5.7 trillion, will be spent on healthcare. 1 These costs create a significant impact to employers, who sponsor most private healthcare plans in the U.S. According to a survey conducted by the National Business Group on Health, the total cost of providing medical and pharmacy benefits for an employee including out-of-pocket costs, premiums, and dependents is expected to average $14,156 per subscriber in 2018. Approximately 70% of this cost is a direct expense to employers, with the remainder borne by employees. This cost has been projected to rise by 5% in 2018, and has risen the past five years consecutively. 2 --- How can the problem be solved? While most employers feel the effects of rising healthcare costs, self-funded employers have an additional, more vested interest in monitoring, influencing and ultimately reducing their healthcare costs, as they bear financial responsibility for most healthcare claims for their employees. Up to 40% of employers in the US are self-insured, according to the Society for Human Resource Management. 3 One option many self-funded employers have turned to is employer-sponsored primary care clinics for employees, including onsite, shared-site and near-site options. In 2015, about 29% of employers with 5,000 or more employees provided an onsite or near-site clinic offering primary care services, according to the National Survey of Employer-Sponsored Health Plans. 4 OurHealth was established with the goal of increasing healthcare accessibility, affordability, and value for employers. Working directly with healthcare plan sponsors, OurHealth delivers primary care services and wellness strategy through fully independent onsite and near-site clinics. Healthcare plan members engage in services through primary care provider, nursing, health coaching and telephonic visits, as well as lab testing and other wellness services. This is performed in a value-based capitated model that does not rely on fee-for-service pricing, allowing OurHealth and its providers to focus on improved engagement and health outcomes rather than transactional costs. In 2015, about 29% of employers with 5,000 or more employees provided an onsite or near-site clinic offering primary care services. For OurHealth clients, the primary goals of investing in the model and its service are to reduce healthcare costs and produce healthier outcomes. This analysis will answer the key question: Do patients who engage with OurHealth employer-based primary care services improve their health and reduce healthcare costs? 4
--- Analysis Methods To answer this question, OurHealth analyzed datasets from three client populations and the respective outcomes in the following categories: health improvements, cost of care, and emergency department utilization. These employers were generally representative of OurHealth's book of business and had multiple years of accurate, complete datasets available, including medical and pharmacy claims, wellness screening results, and clinical data. To ensure the comparison was equivalent, the Johns Hopkins ACG Clinical Grouper was utilized to calculate concurrent risk scores using payer medical and pharmacy claims data from the baseline and analysis periods, with three months of claim run-out (paid dates). Population Overview Each employer must have engaged with OurHealth services for at least two years and provided at least one year of historical claims data prior to gaining access to OurHealth. The analysis includes data produced by clinic interactions from eligible employees, as well as eligible spouses and dependents. Employer 1: A municipal organization in the Southeastern United States with approximately 7,000 eligible employees. Employer 2: A manufacturer in the Midwestern United States with approximately 750 employees. Employer 3: A utility system operations organization with locations in the Midwestern United States with approximately 700 employees. Definition of Engagement Meaningful engagement is defined as a patient visiting an OurHealth provider or health coach three or more times within the measurement period of the past three years. It s important to note that the definition Overall Patient Demographics Employer 1 Employer 2 Employer 3 Gender Breakdown Average Age (Employee & Spouse) 56% M / 44% F 53% M / 47% F 48% M / 52% F 44.1 50.9 45.2 Members 16,694 1,851 2,360 Employee % of Members 42.4% 41.8% 30.1% Employees with 1 Visit (Any visit type) Members with 3 Provider or Health Coach Visits (Engaged) Top Disease State Summary (% of employees) 83.1% 47.6% 84.2% 31.6% 29.6% 38.7% Employer 1 Employer 2 Employer 3 Lipid Disorders 23.1% 24.4% 10.3% Diabetes 9.1% 10.9% 5.5% Hypertension 22.7% 33.7% 13.8% Chronic Obstructive Pulmonary Disease 12.7% 17.5% 9.4% Low Back Pain 13.2% 20.6% 10.6% See Appendix A for more information 5
for non-engaged includes members who may have visited and/or used OurHealth services and experienced health improvements, but did not exceed the three-visit threshold. Health Improvement Method To examine health improvement outcomes, biometric risk values (described below) were compared. The comparison primarily focused on Hemoglobin A1c values but also examined low-density lipoprotein (LDL) and total cholesterol (TC) values. The results set was limited to measurements received within the past three years. Members who interact with OurHealth frequently have multiple visits and lab value screenings. Health outcome improvements were determined by comparing the difference between a patient s initial measurement (T1) for the selected lab value and latest measurement (T2). Additionally, this analysis only includes patients with lab values at least 90 days apart between their initial (T1) and latest (T2) measurements to allow a reasonable period for value changes to occur. To determine the cumulative improvement, the average improvement was then multiplied by the number of members with at least two measurements (T1 and T2) within the cohort for the total improvement from baseline to current. Measurement Description Optimal Range Improvement Correlation Hemoglobin A1c (HbA1c) A measurement of the last 90 days of glycemic control. High HbA1c values are found in diabetics. 5.7 Reduction Low-Density Lipoprotein (LDL) Commonly known as the bad cholesterol. A high value correlates to a greater risk for heart disease. 100 Reduction Total Cholesterol (TC) Total amount of cholesterol in blood. Includes LDL and HDL. A high value correlates to a greater risk for heart disease. 200 Reduction Emergency Department Utilization Method This analysis focused on all causes of Emergency Department (ED) utilization and examined the number of unique claims per 1,000 members, risk-adjusted by year. * The Johns Hopkins ACG concurrent risk score was used for the risk adjustment for each member for each year in the analysis period. This analysis compared engaged vs. non-engaged cohorts year over year to identify trends. Both cohorts were risk-adjusted to an index of 1. * ED claims are defined as unique claims with a revenue code of 450, 451, 452, 456 or 459, or a CPT code of 99281, 99282, 99283, 99284, 99285 or place of service of 23. 6
Cost of Care Method For cost of care analysis, engaged and non-engaged cohorts are compared based on risk-adjusted PMPY incurred cost amounts. Only employee member data was used. Each employee had to be eligible for the OurHealth plan benefit for a full baseline year and continuously eligible throughout the measurement period. The baseline year PMPY was risk-adjusted to an index of 1 then trended year over year using the Milliman Medical Index (MMI). 5 See Appendix B for the MMI inflation rates used per analysis year. To eliminate outliers due to catastrophic injury or illness, high-cost claimants with total medical and drug claims of more than $100,000 per year were excluded. Observed vs. expected ratios were compared for each cohort year over year. Each cohort s trend was then compared to each other and the delta between them is shown in savings dollars and percentage. --- Findings Health Improvements Overall, engaged members saw significant reductions in HbA1c, LDL, and total cholesterol (TC) during the three-year time period. Across all three employers analyzed, engaging with OurHealth services created a cumulative HbA1c reduction of more than 1,250 points. Notably, this included 47.4% of prediabetic patients, 12.1% of patients in the high diabetic range, and 9.9% of those in very high diabetic range experiencing changes that moved their HbA1c values to the optimal range. For the 3,052 members identified as within the prediabetic range, 47.4% experienced health improvements that moved them to an optimal range. HbA1c Ranges at T1: < 5.7 5.7 and < 6.5 6.5 and < 9.0 9.0 Optimal Prediabetic High Very High Patients 5,698 3,052 555 182 % Change from T1 score -0.6% -3.0% -3.4% -16.9% Patients moved to Optimal on T2 score 1,448 67 18 % of category 47.4% 12.1% 9.9% 7
According to the Centers for Disease Control and Prevention (CDC), 1 in 3 American adults have prediabetes and 90% of those with the condition don t know they have it. Approximately 15 to 30% of prediabetics will develop diabetes in the future 6 and the estimated annual costs to cover an employee with diabetes is $7,900 per year. 7 Identifying Prediabetes Financial Impact Prediabetics Identified 3,052 % Progressing to Type 2 * 15% Estimated Annual Cost of Type 2 Diabetes Patient $7,900 Total Potential Cost of Prediabetics advancing to Type 2 $ 3,616,620 % of Prediabetics improved to optimal range 47.4% Total Estimated Savings $1,715,880 Savings per Prediabetic Identified $562.21 *CDC estimate of prediabetics that will advance to Type 2 diabetes Average cost to cover an employee with diabetes, Johns Hopkins Medicine Using the conservative CDC prediabetic to Type 2 diabetic progression rate of 15%, OurHealth s employerbased primary care services create significant projected savings. For the three employers analyzed, the results seen in the prediabetic population are projected to result in an annual savings of $562.21 PMPY for each prediabetic identified and engaged, and $1.7 million in total savings. This analysis also shows that member engagement with OurHealth has led to significant reductions in other values, including: More than 18,200 LDL points reduced, at an average individual reduction of 4.5%. The results seen in the prediabetic population are projected to result in a savings of $562.21 PMPY for each prediabetic identified and engaged, and $1.7 million in total savings. More than 8,800 TC points reduced. High values for these two measurements can indicate risk for conditions such as heart disease, stroke, peripheral arterial disease, Type 2 diabetes, and high blood pressure. 8 8
See Appendix C for full T1-T2 value change tables. The 0 category includes patients for which OurHealth may have received third-party biometric data or those that did not have a provider/health coach visit. Impact of patient engagement on biometric results This section reviews how the number of visits with OurHealth affect overall improvements seen in biometric screening values. The results indicate that using an OurHealth clinic for provider visits leads to improvements. In addition, the effect of consistently engaging in health coaching over time creates a much larger positive impact on average HbA1c levels. The trends clearly indicate more health coaching has a greater impact on reducing HbA1c values. At more than eight visits over three years, these results diminish; however, they still outperform the cohort that did not engage in health coaching. The 5,111 members with no health coaching averaged a 1.8% (-0.101) HbA1c reduction, while the 7,813 members with any amount of health coaching averaged 2.9% (-0.169). That equates to a 67.3% improvement in relative HbA1c values produced by OurHealth health coaching and treatment. Similar results were seen in LDL and TC. That equates to a 67.3% improvement in relative HbA1c values produced by OurHealth health coaching and treatment. Emergency Department Utilization Patients who engage in a meaningful way with OurHealth services reduce their ED utilization. However, this can take up to a year to take effect after beginning services with OurHealth, as it takes time for new members to become engaged. Looking at year-over-year change, there was an average risk-adjusted reduction of 9% in ED visits in the engaged cohort vs. 5% in the non-engaged (Figure 1). When Figure 1 9
comparing the engaged vs. non-engaged rate of riskadjusted ED visits per 1,000, the engaged cohort used the ED 22.8% less than the non-engaged group. This is a substantial difference, resulting in a potential reduction of 435 visits in the engaged group over the two years. At an average ER visit cost of $1,500 9, this reduction results in a projected savings of $652,000 in aggregate. Our analysis found the engaged cohort used the ED 22.8% less than the non-engaged group. Does this effect hold true with smallto medium-sized organizations? Let s take a deeper look at the small- to mid-sized employers (Employer 2 and Employer 3). For these populations, the ED utilization reduction trend was more substantial, suggesting that it takes longer to produce significant outcomes in larger cohorts such as that in Employer 1. As shown in Figure 2, Employer 2 experienced an average reduction of 29.6% of ED visits in the engaged cohort vs. the non-engaged cohort. For the same organization, and limiting the analysis to just employees and spouses, we see similar results. As seen in Figure 3, the engaged cohort reduced ED department utilization by 26% while the non-engaged cohort rose by 21% in Year 2 (2017). Figure 2 Additionally, across all these analyses, the engaged cohort had a higher risk score than the non-engaged. This means that providing clinic services allows more acute and comorbid patients to access primary care. These reductions in ED utilization can add up to substantial cost reduction and avoidance savings. Cost of Care Member populations who engage in employer-based primary care services generated a higher amount of savings over expected trend than those who did not engage. Additionally, these members cost less PMPY on a risk-adjusted basis. Figure 3 10
Across the pool of three clients, 6,131 members met the criteria for analysis of cost described earlier in this white paper. The engaged cohort included 3,096 members who experienced a 0.83 observed vs. expected ratio. Translated, this means that their costs were 17% lower than expected, adjusting for risk and inflation. At the same time, costs in the non-engaged cohort were 3% (2016) to 8% (2017) lower than expected. Simply put, the engaged cohort saved more money. This trend translates to a two-year trend for medical and drug savings of $2.6M over the non-engaged, or an average of $420 PMPY. Looking deeper, medical claims savings were even steeper with a 19% (2016) and 17% (2017) lower than expected medical spend. Medications saw savings, but not at the same rate as medical, with 10% (2016) and 17% (2017) in the engaged vs. 4% and 2% in the non-engaged, as shown in the bar graphs that follow. This trend translates to a two-year trend for medical and drug savings of $2.6M over the non-engaged, or an average of $420 PMPY. Risk-adjusted PMPY costs for members engaged in OurHealth employer-based primary care services trended much lower than expected over the three-year measurement period. 11
--- Patient Success Meet Tyran D., an employee of Employer 1, who had lived with diabetes for more than 20 years before becoming an eligible OurHealth patient in 2016. Diagnosed in 1991, his HbA1c levels were consistently measuring above 12 as of 2017, which had led to multiple ED visits. After a medical consultation with his OurHealth provider following a wellness screening as part of Employer 1 s incentive strategy, Tyran began regular meetings with a health coach and Certified Diabetes Educator, who emphasized lifestyle changes such as regularly monitoring blood sugar, improving nutrition, and reducing consumption of alcohol. As a result of these interventional changes and consistent engagement, in approximately 12 months, Tyran s HbA1c now regularly measures at 7. His medical claim costs were also reduced by 64%, and his diabetes was better controlled at no out-of-pocket expense to him. --- Summary The analysis found that engaging with OurHealth s employer-based primary care clinic model created substantial health improvements and significant cost savings for the three employer populations surveyed. Patients who used OurHealth for primary care experienced greater health improvement and more savings than those that didn t, even when adjusted for baseline health status. In summary, patients who engage with OurHealth improve their health and create healthcare cost savings for employers. To learn more about OurHealth s data analysis and improved outcomes for patients and employers, visit ourhealth.org To learn more about OurHealth s data analysis and improved outcomes for patients and employers, visit ourhealth.org 12
--- Sources 1. Centers for Medicare and Medicaid, National Health Expenditure Projections 2017-2026, Feb. 2016, https://www.cms.gov/research-statistics-data-and-systems/statistics-trends-and-reports/nationalhealthexpenddata/ NationalHealthAccountsProjected.html 2. National Business Group on Health, Large U.S. Employers Project Health Care Benefit Costs to Surpass $14,000 per Employee in 2018, National Business Group on Health Survey Finds, https://www.businessgrouphealth.org/news/nbgh-news/press-releases/press-release-details/?id=334 3. Society for Human Resource Management, 2017 Health Care Benchmarking Report, https://www.shrm.org/hr-today/trends-and-forecasting/research-and-surveys/documents/2017-health-care-benchmarking.pdf 4. Mercer, National Survey of Employer-Sponsored Health Plans, 2015 https://www.mercer.us/our-thinking/healthcare/employers-launch-worksite-clinics-despite-aca-uncertainty.html 5. Milliman, 2017 Milliman Medical Index, May 2017 http://www.milliman.com/insight/periodicals/mmi/2017-milliman-medical-index/ 6. Centers for Disease Control and Prevention, About Diabetes & Type 2 Diabetes, May 2018ii https://www.cdc.gov/diabetes/prevention/prediabetes-type2/index.html 7. Johns Hopkins Medicine, Diabetes Prevention Program, May 2018 https://www.johnshopkinssolutions.com/solution/act2 8. Cleveland Clinic, Cholesterol: High Cholesterol Diseases https://my.clevelandclinic.org/health/articles/11918-cholesterol-high-cholesterol-diseases 9. Consumer Health Ratings, Emergency Room Typical Average Cost of Hospital ED Visit http://consumerhealthratings.com/healthcare_category/emergency-room-typical-average-cost-of-hospital-ed-visit 13
Appendix A - Employer Demographic, Top Conditions Overall Patient Demographics Employer 1 Employer 2 Employer 3 Gender Breakdown 56% M / 44% F 53% M / 47% F 48% M / 52% F Average Age (Total Population) 35.48 38.9 36.65 Members 16,694 1,851 2,360 Employee % of Members 42.4% 41.8% 30.1% Spouse % of Members 18.0% 23.5% 33.5% Dependent % of Members 33.1% 34.5% 36.3% Age Bands (% of members) Employer 1 Employer 2 Employer 3 0 to 4 1.4% 1.7% 0.0% 5 to 17 19.6% 17.0% 12.0% 18 to 34 26.0% 22.2% 33.6% 35 to 44 16.7% 11.1% 21.0% 45 to 54 19.6% 19.6% 19.6% 55 to 64 14.0% 24.8% 11.2% > 65 2.8% 3.7% 2.6% Top Disease State Summary (% of employees) Employer 1 Employer 2 Employer 3 Diabetes 9.1% 10.9% 5.5% Hypertension 22.7% 33.7% 13.8% Lipid Disorders 23.1% 24.4 10.3% COPD 12.7% 17.5% 9.4% Low Back Pain 13.2% 20.6% 10.6% Mental Health 9.6% 10.6% 4.0% 14
Appendix B - Milliman Medical Index Inflation Rate per Year Type Year MMI Inflation Rate Type Year MMI Inflation Rate Medical 2015 6.3% Pharmaceutical 2015 13.6% 2016 5.3% 2017 4.3% 2016 5.3% 2017 8% Appendix C - HbA1c Value Changes by Number of Provider, Health Coaching Visits Provider Visits 0 1-2 3-4 5-6 7-8 9-10 >10 Average T1 value 5.66 5.67 5.7 5.81 5.91 6.08 6.17 Average T2 value 5.57 5.53 5.53 5.7 5.75 5.84 5.95 Average net change -0.093-0.1371-0.1712-0.1134-0.1588-0.2415-0.2212 Patients 3889 2115 1166 792 486 337 702 Health Coaching Visits 0 1-2 3-4 5-6 7-8 9-10 >10 Average T1 value 5.63 5.74 5.95 6.13 6.67 6.47 6.16 Average T2 value 5.53 5.6 5.78 5.92 6.4 6.13 5.87 Patients 5,111 2,610 1,072 301 170 68 155 15