Predicting use of Nurse Care Coordination by Patients in a Health Care Home Catherine E. Vanderboom PhD, RN Clinical Nurse Researcher Mayo Clinic Rochester, MN USA 3 rd Annual ICHNO Conference Chicago, IL, USA 2014 MFMER slide-1
Acknowledgements Team Members Diane E. Holland PhD, RN Jay Mandrekar, PhD Stephanie G. Witwer PhD, RN Vicki L. Hunt, MD Funding Mayo Clinic Rochester, Division of Nursing Research Mayo Clinic Rochester, Primary Care Internal Medicine 2014 MFMER slide-2
Background Meeting the needs of older adults with multiple chronic conditions is a challenge to the health care system Nurse Care Coordinators (NCCs) in Health Care Homes in primary care settings play a critical role on interdisciplinary teams to improve patient outcomes In order to be successful, NCC services should be targeted to appropriate individuals 2014 MFMER slide-3
Background Prediction models based on retrospective data in administrative databases are currently used to identify patients and often perform poorly Patient appraisal of ability and health status is known to influence use of healthcare services Patient s perceived need for care is rarely included in risk models that identify patients for NCC services 2014 MFMER slide-4
Purpose and Aims Purpose: To identify patients who would likely use NCC services using variables identified in the literature and categorized by the conceptual model. Aims: To examine the extent to which specific factors predict the use of NCC services within the Health Care Home 2014 MFMER slide-5
Conceptual Framework Behavioral Model of Health Service Use Predisposing Characteristics Enabling Resources Need for Care Demographics Social Structure Health Beliefs Psychological Characteristics Personal & Family Resources Individual s Perceived & Professionally Evaluated Use of Health Services 2014 MFMER slide-6
Methods Design: Predictive correlational Sample: 178 community-dwelling older adults Multiple chronic conditions English speaking Offered NCC services Setting: Primary Care/Health Care Home, Mayo Clinic, Rochester, MN USA 2014 MFMER slide-7
Methods Predictor Variables: Potential variables identifed from literature Grouped by concepts in the Conceptual Framework Dependent Variable: Use of NCC services (Yes/No) from administrative data base and confirmed in EHR 2014 MFMER slide-8
Methods Procedures Recruitment: Letter and subsequent phone call to convenience sample of patients previously offered NCC services Data collection: Participant interviews Review of EMR and administrative database 2014 MFMER slide-9
Methods Data Analysis: Group differences compared using 2-sample t, Wilcoxon rank sum, chi-square, or Fisher s exact tests Association of candidate predictors summarized from univariate logistic regression models Multivariable logistic regression model using bootstrap sampling approach Model discrimination summarized using AUC and internally validated using bootstrapping 2014 MFMER slide-10
Predictor Variables (1) Predisposing Characteristics Age, Gender Race/Ethnicity Language Education Occupation Self-efficacy to manage Mental Illness Depression, Anxiety Health Distress Cognition, Decision-making ability Enabling Resources Income Health Insurance Regular Source of Care Marital Status Health status of spouse Adult child living within 10 miles Living Situation Thinking about moving Caregiver availability 2014 MFMER slide-11
Predictor Variables (2) Need for Care Variables Patient s Perceived Need Self-rated health Pain Energy/fatigue Functional Ability - ADL/IADL Restricted Activity Days Bed disability Falls Professionally Evaluated Need Extended diagnostic clusters Diagnoses in EMR Medications Hospitalizations ED use NH use ERA Score Tier 2014 MFMER slide-12
Results Sample Characteristics (N =178) Age mean: 73.7 (9.7) years Gender: 113 (63%) female Race/Ethnicity: 99% White/non-Hispanic 130 (73%) used NCC services 48 (27%) refused NCC services No significant group differences in predisposing characteristics (age, gender, ethnicity, etc.) 2014 MFMER slide-13
Group Differences - Enabling Resources Considering moving to housing with services - Yes Health Insurance Medicare + anything except Medicaid Medicare + Medicaid + anything else Private or commercial insurance Use of NCC NO YES N (%) P-value 2 (4) 19 (15) 0.05 39 (81) 0 9 (19) 113 (87) 10 (8) 7 (5) 0.004 No significant group differences in: Marital status; Health of spouse; Adult child living within 10 miles; Living situation; Assistance for a few days if needed; Assistance for extended time if needed; Income; Regular source of primary care; Employment Status 2014 MFMER slide-14
Group Differences - Patient Perceived Need (1) Mean (Median; QR) Use of NCC NO YES P-value Functional Ability IADL services (#) 1.8 (1; 0-4) 2.9 (3; 1-5) 0.001 Any functional assistance services (#) 3.7 (2; 0-5) 6.2 (4; 1-9) 0.002 ADL Independent Dependent IADL Independent Dependent Combination of ADL and IADL Independent for both Independent for ADL, Dependent for IADL Dependent for both Transportation assistance Independent Dependent 42 (88) 6 (13) 22 (46) 26 (54) 22 (46) 20 (42) 6 (13) 33 (69) 15 (31) 90 (69) 40 (31) 25 (19) 105 (81) 25 (19) 65 (50) 40 (31) 63 (48) 67 (52) 0.01 <0.001 <0.001 0.01 Receiving any functional assistance Independent Dependent 21 (44) 27 (56) 21 (16) 109 (84) <0.001 Self-rated health Excellent or very good Good, fair, or poor 11 (23) 117 (90) 13 (10) 37 (77) 0.02 2014 MFMER slide-15
Group Differences - Patient Perceived Need (2) Use of NCC NO YES Mean (Median; QR) P-value Pain 52.9 (55; 41-70) 54.6 (60; 38-74) 0.39 Energy 2.2 (2.3; 1.2-3.1) 2.2 (2.2; 1.2-3.0) 0.72 # Restricted activity days 24.1 (2; 0-40) 33.8 (7; 0-90) 0.08 # Bed disability days 3.4 (0; 0-1) 2.6 (0; 0-0) 0.56 # Falls in past 6 months 1.5 (0; 0-1) 0.8 (0; 0-1) 0.42 Functional Ability 0.6 (0; 0-0) 1.1 (0; 0-2) 0.06 Medication services (#) Procedure/treatment services (#) 0.1 (0; 0-0) 0.3 (0; 0-0) 0.07 Medication assistance Independent Dependent Procedure/treatment assistance Independent Dependent N (%) 38 (79) 10 (21) 44 (92) 4 (8) N (%) 85 (65) 45 (35) 104 (80) 26 (20) 0.07 0.06 2014 MFMER slide-16
Group Differences Professionally Evaluated Need (1) Use of NCC NO YES Mean (Median; QR) P-value # Diagnoses 8.5 (8; 6-10) 11.0 (10; 8-13) <0.001 # Prescription 14.7 (14; 10-19) 18.5 (18; 14-22) <0.001 medications No significant group differences in: # Extended diagnostic clusters; ERA score; # Days in hospital past 6 months; # ED visits in past 6 months 2014 MFMER slide-17
Group Differences Professionally Evaluated Need (2) Use of NCC NO YES N (%) P-value Presence of extended diagnostic cluster (22) Neurologic 19 (40) 78 (62) 0.01 Cardiovascular 36 (77) 86 (69) 0.32 Dental 9 (19) 19 (15) 0.53 Gastrointestinal/hepatic 20 (43) 72 (58) 0.07 Infections 10 (21) 32 (26) 0.56 Malignancies 7 (15) 18 (14) 0.93 Musculoskeletal 39 (83) 100 (80) 0.66 2014 MFMER slide-18
Univariate Associations with use of NCC Predisposing Characteristics Odds Ratio (95% CI) P-value Employment status 2.31 (1.11 4.79) 2 0.03 Depression No Yes - Down, depressed, hopeless Cognition Cognitively intact Cognitively impaired 1.0 (reference) 1.88 (0.89 3.94) 0.09 1.0 (reference) 2.47 (0.90 6.81) 0.08 2014 MFMER slide-19
Univariate Associations with use of NCC Enabling Resources Odds Ratio (95% CI) P-value Health of spouse (N=104) Excellent or very good Good, fair, or poor Considering moving to housing with services No Yes Health Insurance Medicare and/or Medicaid Private or Insurance 1.0 (reference) 0.45 (0.19 1.07) 0.07 1.0 (reference) 3.94 (0.88 17.59) 0.07 1.0 (reference) 0.25 (0.09 0.71) 0.01 Employment status 2.31 (1.11 4.79) 0.03 2014 MFMER slide-20
Univariate Associations with use of NCC Patient Perceived Need Functional Status ADL Independent Dependent IADL Independent Dependent Combination of ADL and IADL Independent for both Independent for ADL, Dependent for IADL Dependent for both Medication Assistance Independent Dependent Procedure/treatment Assistance Independent Dependent Odds Ratio (95% CI) P-value 1.0 (reference) 3.11 (1.22 7.91) 0.02 1.0 (reference) 3.55 (1.74 7.27) <0.001 1.0 (reference) 2.86 (1.34 6.12) 5.87 (2.09 16.46) 0.01 <0.001 1.0 (reference) 2.01 (0.92 4.41) 0.08 1.0 (reference) 2.75 (0.91 8.35) 0.07 Transportation Assistance 1.33 (1.06 1.68) 2 0.02 Self-rated health Excellent or very good Good, fair, or poor 1.0 (reference) 2.68 (1.11 6.48) 0.03 2014 MFMER slide-21
Univariate Associations with use of NCC Professionally Evaluated Need Odds Ratio (95% CI) P-value Number of diagnoses 1.19 (1.07 1.33) 2 0.001 Number of prescription medications 1.12 (1.05 1.19) 2 <0.001 Presence of extended diagnostic cluster Neurologic diagnostic cluster 2.45 (1.23 4.86) 0.01 2014 MFMER slide-22
Multivariable Model to Predict use of NCC Combination of ADL and IADL Independent for both Independent for ADL, Dependent for IADL Dependent for both ADL and IADL Odds Ratio (95% CI) 1.0 (reference) 2.68 (1.22 5.87) 5.30 (1.81 15.52) P-value 0.014 0.002 Number of prescription medications 1.12 (1.04 1.20) 1 0.002 1. Odds ratio represents a 1-unit increase. 2014 MFMER slide-23
Predicted Probabilities of using NCC Based on Multivariable Model Combination of ADL/IADL Number of Prescription Drugs Predicted Probability of Using NCC Services Independent for both ADL and IADL 10 0.37 Independent for both ADL and IADL 15 0.51 Independent for both ADL and IADL 20 0.64 Independent for both ADL and IADL 25 0.76 Independent for ADL, Dependent for IADL 10 0.62 Independent for ADL, Dependent for IADL 15 0.74 Independent for ADL, Dependent for IADL 20 0.83 Independent for ADL, Dependent for IADL 25 0.89 Dependent for both ADL and IADL 10 0.76 Dependent for both ADL and IADL 15 0.85 Dependent for both ADL and IADL 20 0.91 Dependent for both ADL and IADL 25 0.94 2014 MFMER slide-24
1.0 0.9 0.8 Predicted Probabilty of Use of NCC Services 0.7 0.6 0.5 0.4 0.3 0.2 0.1 Dependent for both ADL and IADL Independent for ADL, Dependent for IADL Independent for both ADL and IADL 0.0 10 15 20 25 Number of Prescription Drugs 2014 MFMER slide-25
Discussion ADL and IADL limitations and use of prescription medications were predictive of NCC use - inclusion of these variables may improve targeting of NCC Results support the importance of patients perceived need for care as well as professionally evaluated need Using prediction models that incorporate sociodemographic variables can optimize the use of limited resources and enhance patient outcomes 2014 MFMER slide-26
Limitations Single primary care practice, likely under represents small, rural practices One care delivery model, not applicable to other settings using other models of care Limited racial/ethnic and geographic diversity Holistic perspective of theoretical framework required including large number of variables however other variables may exist 2014 MFMER slide-27
Thank you! Vanderboom.catherine@mayo.edu 2014 MFMER slide-28