CHRONIC disease reaps a considerable

Transcription

1 J Ambulatory Care Manage Vol. 30, No. 1, pp c 2007 Lippincott Williams & Wilkins, Inc. Is Patient Activation Associated With Outcomes of Care for Adults With Chronic Conditions? David M. Mosen, PhD, MPH; Julie Schmittdiel, PhD; Judith Hibbard, PhD; David Sobel, MD, MPH; Carol Remmers, MPH; Jim Bellows, PhD Abstract: We examined the patient activation measure s (PAM s) association with process and health outcomes among adults with chronic conditions. Patients with high PAM scores were significantly more likely to perform self-management behaviors, use self-management services, and report high medication adherence, compared to patients with the lowest PAM scores. This population was 10 times more likely to report high patient-satisfaction scores, 5 times more likely to report high quality-of-life scores, and reported significantly higher physical and mental functional status scores, compared to those with the lowest scores. These results suggest that PAM scores are associated with key process and health outcome measures. Key words: chronic conditions, Patient Activation Measure, quality of care CHRONIC disease reaps a considerable toll on the millions of patients who suffer with them (Wagner et al., 1996). This is complicated by the fact that the prevalence of chronic conditions is increasing (Institue of Medicine [IOM], 2001). This large increase in the prevalence of chronic conditions in the United States places unique demands on the healthcare delivery system (Renders et al., 2001). Unfortunately, medical professionals attitudes were developed at a time predominated by acute, infectious diseases; these attitudes are simply not consistent with what is needed presently to man- From the Kaiser Permanente Center for Health Research, Portland (Dr Mosen); Kaiser Permanente Division of Research, Oakland (Dr Schmittdiel); Department of Planning, Public Policy and Management, University of Oregon (Dr Hibbard); Kaiser Permanente Northern California Regional Health Education (Dr Sobel); and Kaiser Permanente Care Management Institute, Oakland (Ms Remmers and Dr Bellows). Corresponding author: David M. Mosen, PhD, MPH, Kaiser Permanente Center for Health Research, 3800 N Interstate Ave, Portland, OR ( David.M.Mosen@kpchr.org). age chronic conditions (Wagner et al., 1996). Not only are chronic conditions different in their time course and severity, they also require the patient to make nearly continuous decision making owing to constantly changing circumstances (Holman & Lorig, 2000). As a result, new patient-centered measures are needed to assess patients abilities to manage chronic conditions. The Patient Activation Measure (PAM), developed by Hibbard and colleagues (2004, 2005), is one such measure that deserves further investigation. The PAM assesses a person s knowledge, skill, and confidence for managing one s own healthcare. This type of assessment is important because patients are more likely to make good decisions and promote their own health if they are more engaged, informed, and feel confident that they can take care of themselves (Lorig, 1996; Lorig et al., 1999; Von Korff et al., 1997, 1998). Assessment of patient activation using the PAM tool may also be important from a policy perspective. The recent Institute of Medicine Summit on Crossing the Quality Chasm suggested new directions in quality measurement, recommending that 21

2 22 JOURNAL OF AMBULATORY CARE MANAGEMENT/JANUARY MARCH 2007 measurement should focus on the patient and monitor their experiences (IOM, 2001). The IOM recommended that such measurement should be integrated into the care delivery process and improve patient care. Such measurement should also be longitudinal and should capture patients experiences over time. The PAM has the potential to meet these requirements and its use could further illuminate how chronic conditions affect this population (Hibbard et al., 2004). Previous research has found that the PAM is associated with improved outcomes. In a telephone survey of 1515 adults, Hibbard and colleagues (2004) found those with higher activation were more likely to exercise regularly, consume a low-fat diet, eat more fruits and vegetables, not smoke, and engage in consumerist health behaviors, such as finding a new provider s qualifications. No other research to date, however, has examined the PAM s association with comprehensive process measures and health-related outcomes. This article s primary objective is to examine the PAM s association with processand health-related outcomes for adults with chronic health conditions. METHODS Data source In September 2004, we sent surveys to a random sample of Kaiser Permanente (KP) Medical Care program members from 1 of 6 chronic condition populations: (1) asthma, (2) diabetes (DM), (3) heart failure, (4) coronary artery disease (CAD), (5) chronic pain, and (6) both DM and CAD. The study sample was representative of KP s membership population and was selected from 7 of KP s 8 regions: Northern California, Southern California, Colorado, Georgia, Hawaii, Mid-Atlantic States, and Northwest. These regions represent about 90% of KP s total membership. We identified the study population from the Clinical Outcomes Research and Evaluation (CORE) database of the KP Care Management Institute. CORE identifies populations with chronic conditions on the basis of information from KP automated databases for pharmacy, inpatient and outpatient utilization, and laboratory use. Information from CORE is used to produce national evaluation reports (Kaiser Permanente Care Management Institute, 2003) and identify study populations that have been previously reported in peer-reviewed literature (Schatz et al., 2005a, 2005b). This population s overall physical and functional status is similar to that of other published research for populations with chronic conditions (Ware et al., 2001). We sent a total of 8908 surveys to the study population. Surveys were administered in English by mail and telephone. We first sent the survey by mail, and followed this with a telephone contact 3 weeks later for those that did not respond by mail. A total of 12 attempts were made to those administered the survey by telephone. We were able to reach 6673 respondents, yielding a contact rate of 75.4%. Reasons for no contact included incorrect addresses and phone numbers, participant was no longer a KP member, or the participant was not reachable. Of the 6673 respondents contacted, 4108 completed the survey, for a total response rate of 61.2%. Respondents and nonrespondents did not differ on the basis of age, gender, and chronic condition (results not shown). The KP Northern California Institutional Review Board approved the study. Survey elements The cross-sectional survey included questions regarding age, sex, race/ethnicity, education (expressed as a 6-point scale from eighth grade or less to a postgraduate or professional degree), and self-reported height and weight (used to calculate body mass index). The 22-item Patient Activation Measure (PAM) was the primary independent variable used in the present analysis. The PAM was created using Rasch methodology and is a 1-dimensional, interval-level, Guttman-like scale (Hibbard et al., 2004). The PAM has been extensively tested and shown to have strong psychometric properties (Hibbard et al., 2004). It is measured on a theoretical 0 to 100 scale (0 = lowest activation, 100 = highest activation).

3 Patient Activation and Outcomes of Care for Adults With Chronic Conditions 23 The study included 6 dependent variables: 3 process measures and 3 health-related outcome measures. The process measures included use of self-management services for chronic conditions, performance of selfmanagement behaviors, and medication adherence. Use of self-management services were assessed with a set of 5 yes/no questions asking whether the participant used common self-management services available 6 months prior to the interview. Selfmanagement services included use of the following: the Kaiser Permanente Web site, KP health education classes, KP Health- Wise Handbook, KP HealthPhone (prerecorded health education audiotapes), and KP emotional support groups (specific questions listed in the Appendix). Performance of self-management behaviors included 4 items adapted from Hibbard and colleagues (2004) scored on a 4-point Likert-type scale (1 = strongly disagree, 2 = disagree, 3 = agree, 4 = strongly disagree) to create a self-management behavior index (Min = 4, Max = 16; Cronbach α =.63). Behaviors addressed included daily consumption of 5 servings of fruits and vegetables, performing tasks needed to manage chronic condition, following a regular exercise program, and following a regular stress management program (specific questions listed in the Appendix). Participants were asked how many days of medication doses were missed in the 7 days prior to interview to assess medication adherence (specific questions listed in the Appendix). We adapted this measure from an item developed by Chesney and colleagues as part of the Outcomes Committee of the Adults AIDS Clinical Trials (AACTG; Chesney et al., 2000; Wagner et al., 2001). Wagner and colleagues (2001) have validated Chesney s measure of self-reported medication adherence within an HIV population. Specifically, the measure was found to have high internal validity and be predictive of likely markers of nonadherence, such as decreased viral load, current alcohol use, current drug use, and depressive symptoms. The health-related outcome measures included overall satisfaction with care, selfreported quality of life, and functional status. Participants were asked to rate their overall satisfaction with the healthcare and services provided by Kaiser Permanente (0 = worst healthcare possible, 10 = best healthcare possible) in the year prior to their interview to assess satisfaction with care. This measure is currently used in the Consumer Assessment of Health Plans Survey (CAHPS) and has been reported elsewhere (CAHPS 2.0 Survey Reporting Kit, 2002). We assessed quality of life by asking respondents to rate their overall quality of life on a 5-point scale (very good, good, neutral, poor, very poor). This item is currently used in the World Health Organization Quality of Life Brief Survey (WHOQOL-BREF) and has been reported elsewhere (WHOQOL Group, 1998). We assessed functional status using the Short Form-8 (SF-8; Laliberte et al., 2002; Turner-Bowker et al., 2003; Ware et al., 2001) survey to produce the physical and mental component summary scores (PCS and MCS, respectively). Statistical methods Four of the 5 outcome measures were analyzed as a dichotomous measure: use of selfmanagement services (use of 1 or more services vs no use), high medication adherence (missed 1 or fewer medications vs 2 or more medications), high patient satisfaction (satisfaction score of 9 or higher vs 8 or lower; CAHPS Reporting Kit, 2002), and high quality of life (very good/good vs neutral rating/poor/very poor). We analyzed functional status and the self-management performance index as a continuous measure. The PAM measure was analyzed using the 4 stages of activation: Stage 1 does not yet understand an active role is important (score 47.0); Stage 2 lacks knowledge and confidence to take action (score 47.1 and 55.1); Stage 3 beginning to take action (score 55.2 and 67.0); and Stage 4 maintaining behaviors over time (score 67.1). These cutpoints for the 4 stages of activation were empirically derived and were based on unpublished work by Hibbard and colleagues. All analyses were performed using the Statistical Package for the Social Sciences (SPSS 13.0, SPSS Inc, Chicago Ill). We examined the

4 24 JOURNAL OF AMBULATORY CARE MANAGEMENT/JANUARY MARCH 2007 Table 1. Sample characteristics Sample Sample description characteristic (N = 4108) Mean age ± SD 61.9 ± 14.7 Male (%) 50.0 Race/ethnicity (%) White 64.0 African American 12.8 Hispanic 8.6 Asian American 8.1 Other race 4.3 Educational attainment (%) Less than high school 11.0 High school graduation 22.0 Some college/technical 38.7 school College graduation or 28.3 higher Chronic condition cohorts (%) Asthma 13.9 Chronic pain 16.9 Coronary artery disease 17.0 Diabetes 15.5 Heart failure 16.9 Both diabetes and CAD 19.4 Mean body mass index 29.4 ± 6.8 (BMI) ± SD BMI % Geographical Location (%) California regions 67.0 Non-California regions 33.0 PAM measure s bivariate associations with categorical and continuous outcome measures using χ 2 and 1-way ANOVA analysis, respectively. Ordinary least squares and multiple logistic regression models were constructed to analyze the PAM s independent effect on outcome measures adjusting for age, gender, race/ethnicity, chronic condition cohort, BMI, and regional geographic location. RESULTS Sample characteristics Overall, the study population was older (mean age = 60 years), well educated (67% with some college or higher), evenly distributed among men and women, and represented a substantial nonwhite population (36%, Table 1). The sample was representative of Kaiser Permanente s population, with about two thirds of the sample being drawn from the Northern and Southern California Regions. The mean BMI was 29.4 and nearly 40% reported a BMI of 30 or higher, the recognized BMI marker for obesity. The mean PAM score for the population was 56.8 (Table 2). About 12% scored in the Activation Stage 1 range, nearly half scored in the Activation Stage 2 range, and just over one third scored in the Activation Stage 3 or higher. A majority of the study sample reported use of 1 or more self-management services (56.5%) and high medication adherence (90% reported 1 or fewer missed days of medications in the past 7 days). About half the study population reported high satisfaction with overall care and services from Kaiser Permanente (51.3% with satisfaction score 9) and high quality of life (nearly 70% reported very good/good quality of life). Association of PAM with study outcome measures PAM was significantly associated with all study outcomes in both bivariate and multivariable analysis (Tables 3 5). Each increased stage of PAM scores was associated with improved outcomes for performance of self-management behaviors, medication adherence, satisfaction with services, self-reported quality of life, and functional status (Table 3). Overall, higher PAM scores were associated with increased use of selfmanagement services, with the highest use of self-management services reported among those with Stage 3 PAM scores (Table 3), with little difference reported between Stage 3 PAM and Stage 4 PAM scores. These results suggest that PAM is more sensitive in its ability to detect differences in 2 of the 3 process measures studied: performance of self-management behaviors and medication adherence, but less sensitive in its ability to detect differences in use of self-management services.

5 Patient Activation and Outcomes of Care for Adults With Chronic Conditions 25 Table 2. Descriptive statistics: Patient Activation Measure (PAM) and outcome measures Independent variable: Patient Activation Measure Sample size (N = 4108) Mean ± SD (0 = lowest, 100 = highest) 56.8 ± 10.0 Stage 1 PAM (score <47.0) 10.3% Stage 2 PAM (score 47.1 and 55.1) 37.2% Stage 3 PAM (score 55.2 and 67.0) 22.7% Stage 4 PAM (score 67.1) 14.3% Study outcome measures Sample size (N = 4108) Use of Self-Management Services Mean services used ± SD (0 = min, 5 = max) 0.84 ± 0.96 Use of 1 or more self-management services 56.5% Self-Management Behavior Index Mean ± SD (4 = min, 16 = max) 10.6 ± 2.3 Medication Adherence Missed 1 or fewer days of medications in past 7 days 90.1% Satisfaction With Care Mean satisfaction ± SD (0 = mix, 10 = max) 8.2 ± 1.9 High satisfaction ( 9) 51.3% Functional Status Mean Physical Component Summary score ± SD 41.7 ± 11.1 Mean Mental Component Summary score ± SD 48.7 ± 10.2 Quality of Life Rated as very good/good 65.4% Rated as neither good nor poor 23.8% Rated as fair/poor 10.8% Table 3. Association of Patient Activation Measure (PAM) with outcome measures Continuous outcome measures Stage 1 PAM Stage 2 PAM Stage 3 PAM Stage 4 PAM Self-Management Behavior Index Mean ± SD* 9.1 ± ± ± ± 2.4 Functional Status Mean Physical Component 35.6 ± ± ± ± 11.1 Summary score ± SD* Mean Mental Component 43.1 ± ± ± ± 9.1 Summary score ± SD* Categorical outcome measures Utilization of Self-Management Services: Use of 1 service (%)* Medication Adherence: Missed or fewer days of medications in past 7 days (%)* Satisfaction With Services: High satisfaction (% with score 9)* Quality of Life: rated as very good/good (%)* P <.0001.

6 26 JOURNAL OF AMBULATORY CARE MANAGEMENT/JANUARY MARCH 2007 Table 4. OLS Regression results: Independent association of Patient Activation Measure (PAM) with continuous outcome measures (standardized beta coefficients presented) Physical Mental Self-Management Component Component Behavior Summary Summary Independent variables Index score score PAM Stage 1 PAM (reference group) NA NA NA Stage 2 PAM Stage 3 PAM Stage 4 PAM Standardized beta coefficients presented for each model and adjusted for age, gender, educational attainment, race/ethnicity, body mass index status, chronic condition cohort, and regional geographical location. P < With the exception of use of selfmanagement services, the greatest variation in study outcomes was reported between those with Stage 1 (lowest activation) and Stage 4 (highest activation) PAM scores. Compared with those with Stage 1 PAM scores, participants with Stage 4 PAM scores were nearly 3 times more likely to report high medication adherence, over 10 times more likely to report high patient-satisfaction scores, and about 5 times more likely to report high quality-of-life scores. Similarly, those with Stage 4 PAM scores reported 48% higher self-management behavior index scores and 26% and 29% higher PCS and MCS scores, respectively. DISCUSSION This study found that patient activation was independently associated with process measures and health outcome measures for an adult population with chronic conditions. Similar to Hibbard and colleagues (2004), we found that higher patient activation was associated with process measures. Specifically, those with higher PAM scores were more likely to perform self-management behaviors, use self-management services, and report higher medication adherence. Of critical importance, this study was the first to find an independent association between patient activation and health-related outcome measures. Those with the higher PAM scores were significantly more likely to report higher patient satisfaction, higher quality-of-life scores, and higher physical and mental functional status, compared with those with lower PAM scores. We found that those with the lowest PAM scores also reported extremely low PCS scores. In fact, those with the lowest PAM scores reported physical functional status scores over 2 standard deviations below US national norms. These results suggest that PAM may be useful in identifying groups with low physical functional status and in possible need of further care management follow-up. Further work is needed to determine the feasibility of the PAM tool to identify new patients in need of care management services and to determine which specific PAM items are most adaptable in the clinical care setting. Furthermore, we found that the lowest PAM scores were reported among older adults, those with less than a high school education, and those with CAD, identifying other possible target groups for intervention (results not shown). This study had several limitations. First, it is difficult to ascertain causality because of the cross-sectional study design. While

7 Patient Activation and Outcomes of Care for Adults With Chronic Conditions 27 Table 5. Logistic regression: Independent association of Patient Activation Measure (PAM) with categorical outcome measures Use of 1 Missed 1 or fewer High satisfaction: Quality of life rated self-management service days of medications score of 9 very good/good Independent variables OR 95% CI OR 95% CI OR 95% CI OR 95% CI PAM Stage 1 PAM (reference group) 1.00 NA 1.00 NA 1.00 NA 1.00 NA Stage 2 PAM Stage 3 PAM Stage 4 PAM Models adjusted for age, gender, educational attainment, race/ethnicity, body mass index status, chronic condition cohort and regional geographical location. Reference group: no self-management service use. Reference group: missed 2 or more days of medications in past 7 days. Reference group: satisfaction scores of 8 or lower. Reference group: quality of life rated as neither poor nor good or fair or poor. higher PAM scores are associated with improved process measures, such as use of selfmanagement services, it is not clear whether increased patient activation leads patients to engage in more self-management, or whether the process of engaging in self-management activities increases patient activation. It is also not clear whether improved health outcome measures are due to higher PAM scores, or whether the study population had higher levels of PAM scores because of better a priori health status. The link between PAM and these process and outcome measures, however, strengthens the case for the PAM s potential use as the type of patient-focused quality measures recommended by the IOM. While respondents versus nonrespondents in this population were similar, and we garnered a robust response rate, this study did not measure PAM scores in the general KP population without chronic conditions; thus, results are not completely generalizable to the entire KP population. Finally, this study was set in a large group model integrated delivery system. It is possible that results are not generalizable to patients with chronic conditions in other healthcare delivery systems where access to selfmanagement services and prescription medication differs from KP. The KP population, however, has been shown to be representative of the demographics within regions that it serves (Van Den Eeden et al., 2003), which may increase the generalizability of these findings. Further research is needed to examine the PAM s prospective effect on process level measures. Specifically, further investigation research is needed to examine the association of PAM with prospective changes in diseasespecific quality of care and utilization measures. Moreover, more work is needed to examine the impact of incremental changes in PAM scores on subsequent changes in key outcomes. Such information will be useful to national quality oversight organizations, such as the National Committee on Quality Assurance (NCQA), that are interested in identifying patient-centered quality of care measures associated with important outcome measures.

8 28 JOURNAL OF AMBULATORY CARE MANAGEMENT/JANUARY MARCH 2007 REFERENCES CAHPS 2.0 Survey and Reporting Kit. (2002). Silver Spring: MD: Publications Clearinghouse; AHRQ. Chesney, M. A., Ickovics, J. R., Chambers, D. B., Gifford, A. L., Neidig, J., Zwickl, B., et al. (2000). Self-reported adherence to antiretroviral medications among participants in HIV clinical trials: The AACTG adherence instruments. Patient care committee & adherence working group of the outcomes committee of the adult AIDS clinical trials group (AACTG). AIDS Care, 12(3), Hibbard, J. H., Mahoney, E. R., Stockard, J., & Tusler, M. (2005). Development and testing of a short form of the patient activation measure. Health Services Research, 40(6, Pt. 1), Hibbard, J. H., Stockard, J., Mahoney, E. R., & Tusler, M. (2004). Development of the patient activation measure (PAM): Conceptualizing and measuring activation in patients and consumers. Health Services Research, 39(4, Pt. 1), Holman, H., & Lorig, K. (2000). Patients as partners in managing chronic disease. Partnership is a perquisite for effective and efficient health care. British Medical Journal, 320, Institute of Medicine. (2001). Crossing the quality chasm: A new health system for the 21st century. Washington, DC: National Academy Press. Kaiser Permanente Care Management Institute. (2003). Unpublished national outcomes reports: Asthma, diabetes, cardiovascular disease, heart failure, and chronic pain. Oakland, CA: Author. Laliberte, K., Turner-Bowker, D., & Ware, J. (2002). Normbased interpretation guidelines (NBIG) for obese adults: A manual for users of the SF-36, SF-12, & SF-8 health surveys. Lincoln, RI: QualityMetric Inc. Lorig, K. R. (1996). Outcome measures for health education and other health care interventions. Thousand Oaks, CA: Sage. Lorig, K. R., Sobel, D. S., Stewart, A. L., Brown, B. W., Jr.; Bandura, A.; Ritter, P., et al. (1999). Evidence suggesting that a chronic disease self-management program can improve health status while reducing hospitalization: A randomized trial. Medical Care, 37, Renders, C. M., Valk, G. D., Griffin, S. J., Wagner, E. H., Eijk van J. ThM, Assendelft, W. J. J., et al. (2001). Interventions to improve the management of diabetes in primary care, outpatient, and community settings: A systematic review. Diabetes Care, 24, Schatz, M., Mosen. D., Apter, A. J., Zeiger, R. S., Vollmer, W. M., Stibolt, T. B., et al. (2005a). Relationship among quality of life, severity, and control measures in asthma: An evaluation using factor analysis. Journal of Allergy and Clinical Immunology, 115(5), Schatz, M., Mosen, D., Apter, A. J., Zeiger, R. S., Vollmer, W. M., Stibolt, T. B., et al. (2005b). Relationship of validated psychometric tools to subsequent medical utilization for asthma. Journal of Allergy and Clinical Immunology, 115(3), Turner-Bowker, D. M., Bayliss, M. S., Ware, J. E., & Kosinski, M. (2003). Usefulness of the SF-8 TM health survey for comparing the impact of migraine and other conditions. Quality of Life Research, 12, Van Den Eeden, S., Tanner, C. M., Berstein, A. L., Fross, R. D., Leimpeter, A., Bloch, D. A., et al. (2003). Incidence of Parkinson s disease: Variation by age, gender and race/ethnicity. American Journal of Epidemiology, 157(11), Von Korff, M., Gruman, J., Schaefer, J., Curry, S. J., & Wagner, E. H. (1997). Collaborative management of chronic illness. Annals of Internal Medicine, 127, Von Korff, M., Katon, W., Bush, E. H., Lin, E. H., Simon, G. E., Saunders, K., et al. (1998). Treatment costs, cost offset and cost-effectiveness of collaborative management of depression. Psychosomatic Medicine, 60(2), Wagner, E. H., Austin, B. T., & Von Korff, M. (1996). Organizing care for patient with chronic illness. Millbank Quarterly, 74, Wagner, J. H., Justice, A. C., Chesney, M., Sinclair, G., Weissman, S., & Rodriguez-Barradas, M. (2001). Patient- and provider-reported adherence: Toward a clinically useful approach to measuring antiretroviral adherence. Journal of Clinical Epidemiology, 54, S91 S98. Ware, J. E., Kosinski, M., Dewey, J. E., & Gandek, B. (2001). How to score and interpret single-item health status measures: A manual for users of the SF-8 health survey (With a supplement on the SF-6 health survey). Lincoln, RI: QualityMetric, Inc. WHOQOL Group. (1998). Development of the World Health Organization WHOQOL-BREF quality of life assessment. Psychology Medicine, 28(3),

9 Patient Activation and Outcomes of Care for Adults With Chronic Conditions 29 Appendix T1. Example process outcome measures Use of Self-Management Services Question In the past 6 months, have you used any of the following Kaiser Permanente services for your health condition? 1. Kaiser Permanente Web site (eg, KP Online, KP.ORG) 2. Health education classes 3. Emotional support groups 4. KP Healthwise Handbook 5. KP Health Phone for ANY Health Condition (recorded audiotapes)? Performance of Self-Management Behaviors In the past 6 months, please indicate how strongly you agree or disagree with each of the following statements: 1. On most days I ate at least 5 servings of fruits or vegetables 2. I did the different tasks and activities needed to manage my health condition so as to reduce my need to see a doctor 3. I followed a regular exercise schedule such as walking, running, swimming, doing aerobics, or using exercise equipment 4. I followed a regular schedule of doing stress management or relaxation techniques Response choice Yes/No Response choice Disagree strongly disagree/agree/agree strongly Medication Adherence During the past 7 days, (including last weekend), on how many days have you missed taking ANY of your doses? 0day 1day 2days 3days 4days 5days 6days 7days