Patient-Centered Medical Homes and the Health of Ohio s Adults and Children

Transcription

1 Patient-Centered Medical Homes and the Health of Ohio s Adults and Children Thomas Wickizer, Kenneth Steinman, Abigail Shoben, Deena Chisolm, Jeff Biehl, Lauren Phelps #OMAS2015 1

2 Please note: This study examines survey respondents perceptions of their health care and whether they reflect care consistent with a patient-centered medical home (CC-PCMH). It does not assess whether individuals actually received care from a certified or accredited PCMH. 2

3 OUTLINE Highlights Background Methods Findings Conclusions / Implications 3

4 KEY HIGHLIGHTS Lower income adults, whether on Medicaid or employer-sponsored insurance (ESI) have similar access to CC-PCMH Lower income adults and children with CC-PCMH are less likely to have: unmet health needs frequent emergency department visits misused prescription painkillers CC-PCMH is less common for African-Americans vs. whites CC-PCMH equally benefits African-Americans and whites Less benefit for pregnant women 4

5 BACKGROUND 5

6 What is a PCMH? Patient-Centered Medical Home (PCMH): a model of coordinated, comprehensive primary care improve outcomes reduce costs increase patient/provider satisfaction 6

7 PCMH in Ohio PCMH Education Pilot Project led 42 primary care practices through a 2-year transformation process. Comprehensive Primary Care Initiative, with 61 practices in southwest Ohio (sponsored by CMS) $75 million State Innovation Model grant from CMS to develop payment systems that will facilitate PCMH development and practice. 7

8 PCMH in OMAS Respondents perspective Not assessing certified PCMH locations Care consistent with a PCMH CC-PCMH Broad statewide perspective Across all health systems; include the uninsured 8

9 Research Questions Does Medicaid facilitate access to CC-PCMH? Is CC-PCMH associated with better health care? Does CC-PCMH reduce health disparities? 9

10 METHODS 10

11 Sample 2015 Ohio Medicaid Assessment Survey (OMAS) 42,876 adults 10,122 proxy interviews of children Focus on Medicaid-covered and potentially Medicaid-eligible 138% FPL for adults 200% FPL for children 11

12 Defining CC-PCMH (1) Has an appropriate, usual source of care (e.g., doctor s office); (2) Has a personal care provider (PCP; i.e., a health professional who knows you well and is familiar with your health history ); (3) Has seen this PCP in the past 12 months; (4) PCP communicates well; (5) Got urgent care (if needed) on the same/next day; (6) Got after hours care (if needed) without a problem; (7) Got specialist care (if needed) without a problem. Does not have CC-PCMH no to any yes to all CC-PCMH 12

13 Analyses Multivariable logistic regression Adjust for demographic characteristics and health status (special health care needs; history of chronic conditions) Survey estimates, represent all Ohio Statistical significance, p<0.05 Predicted probabilities: predicted (not observed) values from statistical models the estimated percentage of a hypothetical subpopulation predicted to have the outcome, assuming they have otherwise average characteristics 13

14 FINDINGS 14

15 Findings Adults Children Who has CC-PCMH? Does Medicaid facilitate access to CC-PCMH? Is CC-PCMH associated with better health care? Does CC-PCMH reduce health disparities? 15

16 Findings Which adults have CC-PCMH? 16

17 unadjusted % adults with CC-PCMH CC-PCMH varies by income, age and gender 100% 75% 50% 25% 22.9% 26.6% lower income 40.4% 39.9% 35.2% 35.7% 31.8% 45.6% 50.8% State mean = 40.0% 28.0% 20.9% 17.7% 31.4% 29.6% 39.6% 45.0% 38.9% 51.1% 46.8% 56.4% 52.6% 0% male female Household income (%FPL) Adult age group 17

18 %* adults 138% FPL with CC-PCMH CC-PCMH is similarly common in different regions 50% 40% 30% 23.1% 27.4% 25.8% 29.2% 26.2% 24.0% 25.2% 20% 10% 0% Medicaid Managed Care Region *Predicted probabilities based on statistical models that adjust for demographic and other characteristics 18

19 %* adults 138% FPL with CC-PCMH CC-PCMH is similarly common in different types of counties 50% 40% 30% 26.1% 25.7% 25.3% 27.8% 20% 10% 0% urban suburban rural Applachian rural non- Applachian *Predicted probabilities based on statistical models that adjust for demographic and other characteristics. 19

20 Findings Which children have CC-PCMH? 20

21 unadjusted % children with CC-PCMH CC-PCMH varies by income and age 100% 75% lower income 50% 25% 28.5% 32.5% 35.3% 36.6% 38.9% 35.8% 42.5% 41.8% 49.1% State mean = 39.2% 52.2% 41.9% 38.5% 35.8% 0% < Household income (as %FPL) Child age group 21

22 %* children 200% FPL with CC-PCMH CC-PCMH is similarly common in different regions 50% 40% 30% 34.3% 39.5% 35.6% 31.4% 31.7% 35.9% 34.5% 20% 10% 0% Medicaid Managed Care Region *Predicted probabilities based on statistical models that adjust for demographic and other characteristics 22

23 %* children 200% FPL with CC-PCMH CC-PCMH is less common in rural non-appalachian counties vs. those in rural Appalachian counties 50% 40% 34.4% 37.9% 38.8% 32.6% 30% 20% 10% 0% urban suburban rural Appalachian rural non- Appalachian *Predicted probabilities based on statistical models that adjust for demographic and other characteristics. 23

24 Findings Does Medicaid facilitate access to CC-PCMH among adults? 24

25 %* adults 138% FPL with CC-PCMH CC-PCMH is similarly common for lower income adults covered by Medicaid or other types of insurance 50% 40% 30% 29.4% 31.3% 29.4% 27.7% 20% 10% 0% Medicaid Medicare & other govt Employer-Sponsored Private/Other Uninsured 9.0% *Predicted probabilities based on statistical models that adjust for demographic and other characteristics 25

26 Nearly half of lower income adults with CC-PCMH have Medicaid Estimated number of lower income adults with CC-PCMH, by insurance type/status employersponsored, 96,000 private/other, 47,000 uninsured, 20,000 Medicaid, 298,000 Medicare, 176,

27 Findings Which adults have CC-PCMH? Does Medicaid facilitate access to CC-PCMH among children? Is CC-PCMH associated with better health care? Does CC-PCMH reduce health disparities? 27

28 %* children 200% FPL who have CC-PCMH CC-PCMH is less common among lower income children with Medicaid versus those with employer-sponsored insurance 50% 40% 30% 34.7% 42.0% 31.8% 20% 20.8% 10% 0% Medicaid Employer-Sponsored Private/Other Uninsured *Predicted probabilities based on statistical models that adjust for demographic and other characteristics 28

29 %* children 200% FPL with outcome Most components of CC-PCMH are equally common among lower income children with Medicaid versus ESI 100% 75% 95% 96% 91% 92% 95% 95% 90% 92% 80% 85% 50% 25% 0% difference not significant Has an appropriate usual source of care difference not significant Has a personal care provider (PCP; of those who have an appropriate usual source of care) Medicaid difference not significant Has seen PCP in the past 12 months (of those who have a PCP) difference not significant Has good communication with PCP (of those who have seen PCP in past 12 months) Employer-Sponsored *Predicted probabilities based on statistical models that adjust for demographic and other characteristics difference not significant Got specialist care without a problem (of those who needed it) 29

30 %* children 200% FPL with outcome Lower income children with Medicaid are less likely than those with ESI to get needed urgent care or after hours care without a problem 100% 75% 78% 83% 57% 64% 50% 25% 0% Got urgent care same or next day (of those who needed it) Medicaid Employer-Sponsored Got after hours care without a problem (of those who needed it) *Predicted probabilities based on statistical models that adjust for demographic and other characteristics 30

31 Nearly ¾ of lower income children with CC-PCMH have Medicaid Estimated # of lower income children with CC-PCMH, by insurance type/status employersponsored, 98,000 private/other, 19,000 uninsured, 7,000 Medicaid, 323,

32 Findings Is CC-PCMH associated with better health care among adults? M 32

33 %* adults 138% FPL with unmet health needs Lower income Medicaid adults with CC-PCMH are less likely to have unmet health needs 50% 31.0% 25% 19.2% 0% *Predicted probabilities based on statistical models that adjust for demographic and other characteristics. Has CC-PCMH Does not have CC-PCMH 33

34 %* Medicaid adults 138% FPL with frequent (3+) ED visits in the past year Lower income Medicaid adults with CC-PCMH are less likely to have frequent emergency department visits 23.0% 20% 13.2% 10% 9.5% 4.3% 0% *Predicted probabilities based on statistical models that adjust for demographic and other characteristics. lower income Medicaid adults with special health care needs Has CC-PCMH lower income Medicaid adults without special health care needs Does not have CC-PCMH 34

35 %* Medicaid adults 138% FPL misusing prescription painkillers in the past year Lower income Medicaid adults with CC-PCMH (and who have special health care needs) are less likely to misuse prescription painkillers 10% 5% 5.2% 0% *Predicted probabilities based on statistical models that adjust for demographic and other characteristics. 2.1% lower income Medicaid adults with special health care needs Has CC-PCMH 1.0% 1.6% lower income Medicaid adults without special health care needs Does not have CC-PCMH difference not significant 35

36 %* Medicaid adults 138% FPL who have been hospitalized in the past year Lower income Medicaid adults with CC-PCMH (and who have a history of chronic conditions) are less likely to have an overnight hospital stay 50% 30.3% 25% 23.8% 0% *Predicted probabilities based on statistical models that adjust for demographic and other characteristics. lower income Medicaid adults with a history of chronic conditions Has CC-PCMH 12.6% 11.1% difference not significant lower income Medicaid adults without a history of chronic conditions Does not have CC-PCMH 36

37 Findings Is CC-PCMH associated with better health care among children? 37

38 %* children 200% FPL with unmet health needs Lower income children with CC-PCMH are less likely to have unmet health needs 20% 15.2% 10% 8.0% 8.9% 4.5% 0% *Predicted probabilities based on statistical models that adjust for demographic and other characteristics. all lower income Medicaid children Has CC-PCMH lower income Medicaid children with special health care needs Does not have CC-PCMH 38

39 %* Medicaid children 200% FPL with frequent (3+) ED visits in the past year Lower income Medicaid children with CC-PCMH are less likely to have frequent emergency department visits 20% 18.4% 10% 6.7% 10.9% 3.7% 0% *Predicted probabilities based on statistical models that adjust for demographic and other characteristics. all low income Medicaid children Has CC-PCMH low income Medicaid children with special health care needs Does not have CC-PCMH 39

40 %* children 200% FPL who have been hospitalized in the past year CC-PCMH is not associated with having an overnight hospital stay 20% 13.1% 14.1% 10% 0% *Predicted probabilities based on statistical models that adjust for demographic and other characteristics. 3.7% difference not significant 4.0% all low income Medicaid children Has CC-PCMH low income Medicaid children with special health care needs Does not have CC-PCMH difference not significant 40

41 %* Medicaid children 200% FPL having a well-child visit in the past year Lower income Medicaid children with CC-PCMH are more likely to have a well-child visit 100% 75% 90.9% 92.1% 74.9% 83.9% 50% 25% 0% *Predicted probabilities based on statistical models that adjust for demographic and other characteristics. low income Medicaid children WITHOUT special health care needs Has CC-PCMH Does not have CC-PCMH low income Medicaid children with special health care needs 41

42 Findings Does CC-PCMH reduce health disparities among adults? 42

43 %* adults who have CC-PCMH White adults are more likely to have CC-PCMH 50% 40% 42.1% 36.1% 36.6% 39.6% 44.7% 42.4% 37.5% 32.3% 30% 28.4% 20% 10% 0% *Predicted probabilities based on statistical models that adjust for demographic and other characteristics. overall Medicaid ESI White African-American Hispanic 43

44 %* adults with unmet health needs For both white and African-American lower income adults, those with CC-PCMH are less likely to have unmet health needs 50% 40% 30% 20% 20.4% 32.6% 24.4% 37.9% 10% 0% *Predicted probabilities based on statistical models that adjust for demographic and other characteristics. White Has CC-PCMH Does not have CC-PCMH African-American 44

45 %* pregnant women 200% FPL with outcome CC-PCMH is not associated with health care outcomes among lower income pregnant women 50% 40% 43.2% 44.4% 30% 27.7% 20% 10% 0% *Predicted probabilities based on statistical models that adjust for demographic and other characteristics. 17.7% difference not significant has unmet health needs Has CC-PCMH 12.9% 9.9% difference not significant frequent (3+/year) emergency department visits Does not have CC-PCMH difference not significant overnight hospital stay 45

46 Findings Does CC-PCMH reduce health disparities among children? 46

47 %* children who have CC-PCMH White children are more likely than African-American children to have CC-PCMH 50% 40% 42.6% 37.6% 39.8% 34.8% 46.4% 41.3% 30% 20% 10% 0% *Predicted probabilities based on statistical models that adjust for demographic and other characteristics. overall Medicaid ESI White African-American 47

48 %* infants 200% FPL with outcome For infants from lower income homes, those with CC-PCMH are less likely to have frequent emergency department visits, but are just as likely to have an overnight hospital stay 50% 40% 30% 20% 24.3% 20.9% 10% 0% *Predicted probabilities based on statistical models that adjust for demographic and other characteristics. 6.8% 3.7% frequent (3+/year) emergency department visits Has CC-PCMH Does not have CC-PCMH difference not significant overnight hospital stay 48

49 Summary of Findings Adults Children Who has CC-PCMH? Does Medicaid facilitate access to CC-PCMH? Is CC-PCMH associated with better health care? Does CC-PCMH reduce racial/ethnic disparities? older, female, higher income younger, higher income = ESI; > uninsured < ESI; > uninsured strong, consistent associations White > Afr-Am; CC-PCMH helps both; weak/no effect for pregnant women strong, consistent associations White > Afr-Am; CC-PCMH helps both; weak effect for infants 49

50 CONCLUSIONS/IMPLICATIONS 50

51 Conclusions Medicaid is facilitating access to CC-PCMH CC-PCMH has robust associations with favorable health care outcomes CC-PCMH may reduce certain health disparities 51

52 Policy Considerations Medicaid expansion may have increased access to CC-PCMH Improving access to urgent and after hours care may help facilitate Medicaid children s access to CC-PCMH Care delivery models that promote CC-PCMH may improve health care outcomes Promoting CC-PCMH may help reduce health disparities OMAS can help monitor CC-PCMH 52

53 Thank You 53

54 Questions? 54