The Effectiveness of Patient Activation Measures in Predicting Chronic Disease Self-Management Success

Introduction

Successful long-term management of chronic disease depends heavily on a patient’s ongoing engagement in day-to-day self-care behaviors, including medication adherence, symptom monitoring, dietary modification, and appropriate use of healthcare services between clinical encounters. The Patient Activation Measure (PAM), developed by Hibbard and colleagues, was designed to quantify an individual’s knowledge, skill, and confidence for managing their own health and healthcare, and classifies respondents into one of four activation levels, ranging from Level 1, in which patients tend to be passive and may not yet believe their own actions matter to their health, through Level 4, in which patients have generally adopted the behaviors needed to maintain function and manage their condition, even under stress (Hibbard et al., 2004).

Since its initial validation, the PAM has been widely adopted across primary care, specialty clinics, and population health management programs as a brief, low-burden instrument that can be administered at the point of care and used to tailor the intensity of self-management support offered to an individual patient. A substantial cross-sectional literature has demonstrated that PAM score correlates with numerous favorable health behaviors and outcomes, including medication adherence, preventive service utilization, and self-reported health status (Greene & Hibbard, 2012). However, cross-sectional correlation does not establish that a single baseline PAM score administered at the start of a care relationship can meaningfully forecast which patients will go on to achieve self-management success over a clinically relevant follow-up period, information that would be considerably more useful to clinicians and care teams seeking to allocate limited self-management support resources toward the patients most likely to benefit.

A smaller body of prospective research has examined the predictive validity of baseline PAM score for subsequent outcomes, generally within single-condition populations such as diabetes or heart failure, and has reported generally favorable, though not uniformly consistent, predictive relationships (Mosen et al., 2007; Rask et al., 2009). Fewer studies have examined this question within a mixed chronic-disease population receiving care in routine ambulatory practice, where comorbidity, variation in condition-specific self-management demands, and heterogeneous clinical severity may all influence the strength of the relationship between baseline activation and eventual self-management success. The purpose of this prospective cohort study was to examine whether baseline PAM score and activation level independently predicted a composite measure of chronic disease self-management success at twelve-month follow-up among adults with one or more of four common chronic conditions receiving care in ambulatory practice, and to characterize the magnitude of this predictive relationship across activation levels.

Methods

Design and setting. This study used a prospective, observational cohort design with a twelve-month follow-up period. Participants were recruited from three hospital-affiliated ambulatory care clinics providing primary and specialty chronic disease management services within a single regional health system between March 2023 and February 2024.

Participants. Adults aged 18 years or older with a physician-confirmed diagnosis of type 2 diabetes, essential hypertension, chronic obstructive pulmonary disease, or heart failure, and at least one prior clinic visit for management of the qualifying condition, were eligible for enrollment. Patients with a diagnosis of moderate to severe cognitive impairment, an anticipated life expectancy of less than twelve months, or insufficient English or Spanish proficiency to complete study instruments without an interpreter were excluded. Eligible patients were approached consecutively during a routine clinic visit, and written informed consent was obtained prior to enrollment.

Measures. Patient activation was assessed at baseline using the 13-item PAM (PAM-13), a validated, Guttman-like instrument yielding a raw score converted to a 0–100 activation score and a corresponding activation level from 1 (lowest) to 4 (highest) (Hibbard et al., 2005). The primary outcome, chronic disease self-management success, was assessed at twelve-month follow-up using a pre-specified composite requiring a participant to meet all three of the following criteria: (a) medication adherence of 80% or greater on the Medication Adherence Report Scale, (b) attainment of the condition-specific clinical target defined for their qualifying condition, including hemoglobin A1c below 8.0% for diabetes, blood pressure below 140/90 mmHg for hypertension, absence of a COPD-related hospitalization or emergency department visit in the preceding six months, and New York Heart Association functional class I or II for heart failure, and (c) a self-management behavior score at or above the cohort median on a condition-appropriate behavioral instrument, most commonly the Summary of Diabetes Self-Care Activities or an analogous condition-specific behavioral scale. Covariates, including age, sex, educational attainment, number of concurrent chronic conditions, and a baseline clinical severity index derived from electronic health record data, were collected at enrollment.

Statistical analysis. Baseline characteristics were summarized using descriptive statistics. The association between baseline PAM score, modeled continuously, and self-management success at twelve months was examined using multivariable logistic regression, adjusted for age, sex, education, number of chronic conditions, and baseline clinical severity. A second model examined baseline PAM level, modeled categorically with Level 1 as the reference category, to characterize the shape of the dose-response relationship across activation levels. Model discrimination was assessed using the area under the receiver operating characteristic curve (AUC). Missing outcome data among participants lost to follow-up were addressed using multiple imputation under a missing-at-random assumption, with results compared against a complete-case sensitivity analysis. A two-sided p value of less than .05 was considered statistically significant.

Results

Of 412 enrolled participants, 358 (86.9%) completed the twelve-month follow-up assessment; the remaining 54 participants were lost to follow-up due to withdrawal, relocation, or non-response, and results from imputed and complete-case analyses did not differ meaningfully. Baseline characteristics of the enrolled cohort are summarized in Table 1. The cohort was distributed across all four PAM activation levels at baseline, with the largest proportion of participants classified at Level 3 (32.5%) and the smallest proportion classified at Level 1 (17.5%).

Overall, chronic disease self-management success at twelve months was achieved by 187 of 358 participants with complete follow-up data (52.2%). Achievement of self-management success varied substantially and monotonically across baseline PAM level, as illustrated in Figure 1. Among participants classified at Level 1 at baseline, only 26.4% achieved self-management success at twelve months, compared with 38.9% at Level 2, 61.7% at Level 3, and 71.8% at Level 4, representing an absolute difference of 45.4 percentage points between the lowest and highest activation levels.

In multivariable logistic regression adjusted for age, sex, education, number of chronic conditions, and baseline clinical severity, baseline PAM score, modeled continuously, remained independently associated with self-management success at twelve months (adjusted odds ratio [aOR] 1.04 per point, 95% CI 1.03–1.06, p < .001), indicating that each ten-point increase in baseline PAM score was associated with approximately a 48% increase in the odds of achieving self-management success. Table 2 presents the adjusted odds ratios for self-management success by baseline PAM level relative to Level 1. The adjusted odds of success increased in a stepwise fashion across activation levels, with participants in Level 4 demonstrating more than seven times the adjusted odds of success observed among participants in Level 1.

Of the covariates included in the adjusted model, only number of chronic conditions and baseline clinical severity index reached independent statistical significance alongside PAM score, with each additional chronic condition and each unit increase in severity index associated with modestly reduced odds of self-management success. Age, sex, and educational attainment were not independently associated with the outcome once baseline PAM score was included in the model, suggesting that activation captured much of the variance in self-management success otherwise attributable to these demographic factors. The full adjusted model incorporating baseline PAM level demonstrated good discrimination, with an area under the receiver operating characteristic curve of 0.78 (95% CI 0.73–0.83), compared with an AUC of 0.64 for a reduced model containing only demographic and clinical covariates without PAM level, representing a statistically significant improvement in discrimination (p < .001, DeLong’s test).

Exploratory analysis of the three individual components of the self-management success composite indicated that the strength of the association with baseline PAM level was broadly consistent, though not identical, across components. The association was strongest for the behavioral self-management component and somewhat weaker, though still statistically significant, for attainment of condition-specific clinical targets, consistent with the expectation that clinical target attainment is influenced by additional factors, including disease severity and physiologic response to treatment, that are not fully captured by patient activation alone.

Discussion

This prospective cohort study found that baseline Patient Activation Measure score and level independently and meaningfully predicted chronic disease self-management success at twelve-month follow-up in a mixed population of adults with diabetes, hypertension, heart failure, and chronic obstructive pulmonary disease. The magnitude of the observed relationship, a 45.4 percentage-point absolute difference in success rate between the lowest and highest activation levels, and more than sevenfold higher adjusted odds of success among the most highly activated participants, substantially exceeds what would be expected from a measure with only modest predictive utility, and supports the PAM’s value as a genuine prospective predictor rather than merely a cross-sectional correlate of self-management outcomes.

The stepwise, dose-response pattern observed across activation levels is consistent with the developmental model underlying the PAM instrument itself, in which progression through activation levels is understood to reflect an accumulating sequence of belief, knowledge, skill, and behavioral change rather than a simple linear scale (Hibbard & Cunningham, 2008). The comparatively modest and non-significant difference observed between Level 1 and Level 2 in the adjusted model, contrasted with the much larger jump in adjusted odds between Level 2 and Level 3, may reflect a meaningful clinical threshold, consistent with prior PAM literature suggesting that the transition from Level 2 to Level 3 often represents the point at which patients begin taking independent action on their own health rather than remaining primarily reliant on external prompting (Hibbard et al., 2007).

The finding that baseline PAM score substantially attenuated the independent association between demographic factors, including age, sex, and education, and self-management success has a practical implication for care teams: activation level may offer a more direct and actionable target for intervention than demographic risk factors that are not themselves modifiable. Where a demographic characteristic such as lower educational attainment might traditionally prompt a clinician to assume elevated risk of self-management difficulty, this study’s findings suggest that directly assessing activation may provide a more precise and individualized basis for that risk assessment, and, more importantly, a target that can plausibly be improved through activation-focused coaching interventions, several of which have themselves demonstrated the ability to shift patients upward across PAM activation levels (Hibbard et al., 2007).

These findings support the practical use of the PAM as a brief, point-of-care risk-stratification tool to guide the allocation of self-management support resources. Rather than applying a uniform level of self-management coaching intensity across an entire chronic disease population, care teams could use baseline PAM score to identify patients at Level 1 or Level 2, who in this cohort demonstrated the lowest probability of unaided self-management success, and preferentially direct more intensive, tailored coaching resources toward this subgroup, while patients already functioning at Level 3 or Level 4 may require comparatively lighter-touch support to maintain existing self-management behaviors (Hibbard & Greene, 2013).

Several limitations should be considered when interpreting these findings. This study was conducted within a single regional health system across three affiliated clinics, and generalizability to other care settings, healthcare systems, or populations with different demographic or socioeconomic profiles may be limited. The composite self-management success outcome, while designed to integrate adherence, clinical target attainment, and behavioral measures into a single clinically meaningful endpoint, necessarily involved judgment in the selection of component thresholds, and alternative composite definitions might yield somewhat different effect estimates. Loss to follow-up, while modest at 13.1%, may have introduced bias if attrition was differentially associated with both baseline activation and eventual self-management success, although sensitivity analyses using multiple imputation did not suggest a substantially different pattern of results. Finally, as an observational cohort design, this study can establish predictive association but cannot establish that baseline activation causally determines self-management success; unmeasured confounding, including health literacy, social support, and psychological factors such as self-efficacy or depressive symptoms, may partially account for the observed relationship.

Future research should examine whether activation-focused coaching interventions delivered specifically to patients identified as Level 1 or Level 2 through point-of-care PAM screening produce measurable improvement in subsequent self-management success, which would provide more direct evidence for the actionable, intervention-guiding value of PAM-based risk stratification suggested by this study’s findings. Replication in more diverse healthcare settings and populations, alongside longer follow-up periods extending beyond twelve months, would further strengthen confidence in the durability and generalizability of the predictive relationship identified here. Taken together, this study’s findings support baseline patient activation as a robust, independent predictor of twelve-month chronic disease self-management success and reinforce the clinical utility of incorporating brief PAM assessment into routine chronic disease management workflows as a practical basis for individualizing self-management support.

References

Greene, J., & Hibbard, J. H. (2012). Why does patient activation matter? An examination of the relationships between patient activation and health-related outcomes. Journal of General Internal Medicine, 27(5), 520–526.

Hibbard, J. H., & Cunningham, P. J. (2008). How engaged are consumers in their health and health care, and why does it matter? Center for Studying Health System Change Research Brief, 8, 1–9.

Hibbard, J. H., & Greene, J. (2013). What the evidence shows about patient activation: Better health outcomes and care experiences, fewer data on costs. Health Affairs, 32(2), 207–214.

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), 1918–1930.

Hibbard, J. H., Mahoney, E. R., Stock, R., & Tusler, M. (2007). Do increases in patient activation result in improved self-management behaviors? Health Services Research, 42(4), 1443–1463.

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(4p1), 1005–1026.

Mosen, D. M., Schmittdiel, J., Hibbard, J., Sobel, D., Remmers, C., & Bellows, J. (2007). Is patient activation associated with outcomes of care for adults with chronic conditions? Journal of Ambulatory Care Management, 30(1), 21–29.

Rask, K. J., Ziemer, D. C., Kohler, S. A., Hawley, J. N., Arinde, F. J., & Barnes, C. S. (2009). Patient activation is associated with healthy behaviors and ease in managing diabetes in an indigent population. Diabetes Educator, 35(4), 622–630.

Skolasky, R. L., Mackenzie, E. J., Wegener, S. T., & Riley, L. H. (2011). Patient activation and adherence to physical therapy in persons undergoing spine surgery. Spine, 36(11), E748–E754.

Vukovic, V., Lillini, R., & Federici, A. (2021). Patient activation and health literacy: A systematic review of their relationship with chronic disease self-management outcomes. Patient Education and Counseling, 104(9), 2192–2202.

◆