Association Between Patient Activation and Health Care Utilization After Thoracic and Abdominal Surgery | Surgery | JAMA Surgery | JAMA Network
[Skip to Navigation]
Sign In
Figure.  Patient Selection Flow Diagram
Patient Selection Flow Diagram

ICU indicates intensive care unit.

Table 1.  Baseline Demographic and Clinical Characteristics and In-Hospital Health State
Baseline Demographic and Clinical Characteristics and In-Hospital Health State
Table 2.  Postoperative Clinical and Self-reported Outcomes 30 Days After Hospital Discharge
Postoperative Clinical and Self-reported Outcomes 30 Days After Hospital Discharge
Table 3.  Multivariate Logistic Regression of the Association of Overall Unplanned Health Care Visits With ED Visits 30 Days After Discharge
Multivariate Logistic Regression of the Association of Overall Unplanned Health Care Visits With ED Visits 30 Days After Discharge
Table 4.  Multivariate Logistic Regression of the Association of Variables With 30-Day Overall Complications
Multivariate Logistic Regression of the Association of Variables With 30-Day Overall Complications
1.
Hsia  RY, Niedzwiecki  M.  Avoidable emergency department visits: a starting point.   Int J Qual Health Care. 2017;29(5):642-645. doi:10.1093/intqhc/mzx081PubMedGoogle Scholar
2.
All-Cause Readmission to Acute Care and Return to the Emergency Department. Canadian Institute for Health Information; 2012. https://secure.cihi.ca/free_products/Readmission_to_acutecare_en.pdf
3.
Thanh  NX, Chuck  AW, Wasylak  T,  et al.  An economic evaluation of the Enhanced Recovery After Surgery (ERAS) multisite implementation program for colorectal surgery in Alberta.   Can J Surg. 2016;59(6):415-421. doi:10.1503/cjs.006716PubMedGoogle Scholar
4.
Gillis  C, Gill  M, Marlett  N,  et al.  Patients as partners in Enhanced Recovery After Surgery: a qualitative patient-led study.   BMJ Open. 2017;7(6):e017002. doi:10.1136/bmjopen-2017-017002PubMedGoogle Scholar
5.
Fearon  KC, Ljungqvist  O, Von Meyenfeldt  M,  et al.  Enhanced recovery after surgery: a consensus review of clinical care for patients undergoing colonic resection.   Clin Nutr. 2005;24(3):466-477. doi:10.1016/j.clnu.2005.02.002PubMedGoogle Scholar
6.
Glasgow  RE, Orleans  CT, Wagner  EH.  Does the chronic care model serve also as a template for improving prevention?   Milbank Q. 2001;79(4):579-612, iv-v. iv-v. doi:10.1111/1468-0009.00222PubMedGoogle Scholar
7.
Hibbard  JH, Stockard  J, Mahoney  ER, Tusler  M.  Development of the Patient Activation Measure (PAM): conceptualizing and measuring activation in patients and consumers.   Health Serv Res. 2004;39(4 Pt 1):1005-1026. doi:10.1111/j.1475-6773.2004.00269.xPubMedGoogle Scholar
8.
Hibbard  JH, Mahoney  ER, Stock  R, Tusler  M.  Do increases in patient activation result in improved self-management behaviors?   Health Serv Res. 2007;42(4):1443-1463. doi:10.1111/j.1475-6773.2006.00669.xPubMedGoogle Scholar
9.
Remmers  C, Hibbard  J, Mosen  DM, Wagenfield  M, Hoye  RE, Jones  C.  Is patient activation associated with future health outcomes and healthcare utilization among patients with diabetes?   J Ambul Care Manage. 2009;32(4):320-327. doi:10.1097/JAC.0b013e3181ba6e77PubMedGoogle Scholar
10.
Greene  J, Hibbard  JH.  Why does patient activation matter? an examination of the relationships between patient activation and health-related outcomes.   J Gen Intern Med. 2012;27(5):520-526. doi:10.1007/s11606-011-1931-2PubMedGoogle Scholar
11.
Hibbard  JH, Greene  J, Shi  Y, Mittler  J, Scanlon  D.  Taking the long view: how well do patient activation scores predict outcomes four years later?   Med Care Res Rev. 2015;72(3):324-337. doi:10.1177/1077558715573871PubMedGoogle Scholar
12.
Greene  J, Hibbard  JH, Alvarez  C, Overton  V.  Supporting patient behavior change: approaches used by primary care clinicians whose patients have an increase in activation levels.   Ann Fam Med. 2016;14(2):148-154. doi:10.1370/afm.1904PubMedGoogle Scholar
13.
Hibbard  JH, Greene  J, Tusler  M.  Improving the outcomes of disease management by tailoring care to the patient’s level of activation.   Am J Manag Care. 2009;15(6):353-360.PubMedGoogle Scholar
14.
Mitchell  SE, Gardiner  PM, Sadikova  E,  et al.  Patient activation and 30-day post-discharge hospital utilization.   J Gen Intern Med. 2014;29(2):349-355. doi:10.1007/s11606-013-2647-2PubMedGoogle Scholar
15.
Harvey  L, Fowles  JB, Xi  M, Terry  P.  When activation changes, what else changes? the relationship between change in Patient Activation Measure (PAM) and employees’ health status and health behaviors.   Patient Educ Couns. 2012;88(2):338-343. doi:10.1016/j.pec.2012.02.005PubMedGoogle Scholar
16.
Vandenbroucke  JP, von Elm  E, Altman  DG,  et al; STROBE Initiative.  Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): explanation and elaboration.   Int J Surg. 2014;12(12):1500-1524. doi:10.1016/j.ijsu.2014.07.014PubMedGoogle Scholar
17.
Charlson  M, Szatrowski  TP, Peterson  J, Gold  J.  Validation of a combined comorbidity index.   J Clin Epidemiol. 1994;47(11):1245-1251. doi:10.1016/0895-4356(94)90129-5PubMedGoogle Scholar
18.
Hibbard  JH, Mahoney  ER, Stockard  J, Tusler  M.  Development and testing of a short form of the Patient Activation Measure.   Health Serv Res. 2005;40(6 Pt 1):1918-1930. doi:10.1111/j.1475-6773.2005.00438.xPubMedGoogle Scholar
19.
Devlin  NJ, Krabbe  PF.  The development of new research methods for the valuation of EQ-5D-5L.   Eur J Health Econ. 2013;14(suppl 1):S1-S3. doi:10.1007/s10198-013-0502-3PubMedGoogle Scholar
20.
Myles  PS, Myles  DB, Galagher  W,  et al.  Measuring acute postoperative pain using the visual analog scale: the minimal clinically important difference and patient acceptable symptom state.   Br J Anaesth. 2017;118(3):424-429. doi:10.1093/bja/aew466PubMedGoogle Scholar
21.
Walters  SJ, Brazier  JE.  Comparison of the minimally important difference for two health state utility measures: EQ-5D and SF-6D.   Qual Life Res. 2005;14(6):1523-1532. doi:10.1007/s11136-004-7713-0PubMedGoogle Scholar
22.
Clavien  PA, Barkun  J, de Oliveira  ML,  et al.  The Clavien-Dindo classification of surgical complications: five-year experience.   Ann Surg. 2009;250(2):187-196. doi:10.1097/SLA.0b013e3181b13ca2PubMedGoogle Scholar
23.
Slankamenac  K, Nederlof  N, Pessaux  P,  et al.  The comprehensive complication index: a novel and more sensitive endpoint for assessing outcome and reducing sample size in randomized controlled trials.   Ann Surg. 2014;260(5):757-762. doi:10.1097/SLA.0000000000000948PubMedGoogle Scholar
24.
Dumitra  TC, Trepanier  M, Fiore  JF  Jr,  et al.  The relationship of two postoperative complication grading schemas with postoperative quality of life after elective colorectal surgery.   Surgery. 2019;166(4):663-669. doi:10.1016/j.surg.2019.05.058PubMedGoogle Scholar
25.
Rubin  DB, Schenker  N.  Multiple imputation in health-care databases: an overview and some applications.   Stat Med. 1991;10(4):585-598. doi:10.1002/sim.4780100410PubMedGoogle Scholar
26.
Lateef  F.  Patient expectations and the paradigm shift of care in emergency medicine.   J Emerg Trauma Shock. 2011;4(2):163-167. doi:10.4103/0974-2700.82199PubMedGoogle Scholar
27.
Sabbatini  AK, Kocher  KE, Basu  A, Hsia  RY.  In-hospital outcomes and costs among patients hospitalized during a return visit to the emergency department.   JAMA. 2016;315(7):663-671. doi:10.1001/jama.2016.0649PubMedGoogle Scholar
28.
Greene  J, Hibbard  JH, Sacks  R, Overton  V, Parrotta  CD.  When patient activation levels change, health outcomes and costs change, too.   Health Aff (Millwood). 2015;34(3):431-437. doi:10.1377/hlthaff.2014.0452PubMedGoogle Scholar
29.
Parchman  ML, Zeber  JE, Palmer  RF.  Participatory decision making, patient activation, medication adherence, and intermediate clinical outcomes in type 2 diabetes: a STARNet study.   Ann Fam Med. 2010;8(5):410-417. doi:10.1370/afm.1161PubMedGoogle Scholar
30.
Schmaderer  MS, Zimmerman  L, Hertzog  M, Pozehl  B, Paulman  A.  Correlates of patient activation and acute care utilization among multimorbid patients.   West J Nurs Res. 2016;38(10):1335-1353. doi:10.1177/0193945916651264PubMedGoogle Scholar
31.
Skolasky  RL, Mackenzie  EJ, Wegener  ST, Riley  LH  III.  Patient activation and adherence to physical therapy in persons undergoing spine surgery.   Spine (Phila Pa 1976). 2008;33(21):E784-E791. doi:10.1097/BRS.0b013e31818027f1PubMedGoogle Scholar
32.
Skolasky  RL, Mackenzie  EJ, Riley  LH  III, Wegener  ST.  Psychometric properties of the Patient Activation Measure among individuals presenting for elective lumbar spine surgery.   Qual Life Res. 2009;18(10):1357-1366. doi:10.1007/s11136-009-9549-0PubMedGoogle Scholar
33.
Skolasky  RL, Mackenzie  EJ, Wegener  ST, Riley  LH.  Patient activation and functional recovery in persons undergoing spine surgery.   Orthopedics. 2011;34(11):888. doi:10.3928/01477447-20110922-04PubMedGoogle Scholar
34.
Andrawis  J, Akhavan  S, Chan  V, Lehil  M, Pong  D, Bozic  KJ.  Higher preoperative patient activation associated with better patient-reported outcomes after total joint arthroplasty.   Clin Orthop Relat Res. 2015;473(8):2688-2697. doi:10.1007/s11999-015-4247-4PubMedGoogle Scholar
35.
Brennan  JJ, Chan  TC, Killeen  JP, Castillo  EM.  Inpatient readmissions and emergency department visits within 30 days of a hospital admission.   West J Emerg Med. 2015;16(7):1025-1029. doi:10.5811/westjem.2015.8.26157PubMedGoogle Scholar
36.
Sources of Potentially Avoidable Emergency Department Visits. Canadian Institute for Health Information; 2014. https://secure.cihi.ca/free_products/ED_Report_ForWeb_EN_Final.pdf https://secure.cihi.ca/free_products/ED_Report_ForWeb_EN_Final.pdf2014
37.
Hibbard  JH, Greene  J, Sacks  RM, Overton  V, Parrotta  C.  Improving population health management strategies: identifying patients who are more likely to be users of avoidable costly care and those more likely to develop a new chronic disease.   Health Serv Res. 2017;52(4):1297-1309. doi:10.1111/1475-6773.12545PubMedGoogle Scholar
38.
Lash  RS, Bell  JF, Reed  SC,  et al.  A systematic review of emergency department use among cancer patients.   Cancer Nurs. 2017;40(2):135-144. doi:10.1097/NCC.0000000000000360PubMedGoogle Scholar
39.
Gustafsson  UO, Scott  MJ, Schwenk  W,  et al; Enhanced Recovery After Surgery (ERAS) Society, for Perioperative Care; European Society for Clinical Nutrition and Metabolism (ESPEN); International Association for Surgical Metabolism and Nutrition (IASMEN).  Guidelines for perioperative care in elective colonic surgery: Enhanced Recovery After Surgery (ERAS) Society recommendations.   World J Surg. 2013;37(2):259-284. doi:10.1007/s00268-012-1772-0PubMedGoogle Scholar
40.
Wood  T, Aarts  MA, Okrainec  A,  et al; iERAS group.  Emergency room visits and readmissions following Implementation of an Enhanced Recovery After Surgery (iERAS) program.   J Gastrointest Surg. 2018;22(2):259-266. doi:10.1007/s11605-017-3555-2PubMedGoogle Scholar
41.
Gleason-Comstock  J, Streater  A, Ager  J,  et al.  Patient education and follow-up as an intervention for hypertensive patients discharged from an emergency department: a randomized control trial study protocol.   BMC Emerg Med. 2015;15:38. doi:10.1186/s12873-015-0052-3PubMedGoogle Scholar
42.
Yun  PS, MacDonald  CL, Orne  J,  et al.  A Novel surgical patient engagement model: a qualitative study of postoperative patients.   J Surg Res. 2020;248:82-89. doi:10.1016/j.jss.2019.11.025PubMedGoogle Scholar
43.
Blakemore  A, Hann  M, Howells  K,  et al.  Patient activation in older people with long-term conditions and multimorbidity: correlates and change in a cohort study in the United Kingdom.   BMC Health Serv Res. 2016;16(1):582. doi:10.1186/s12913-016-1843-2PubMedGoogle Scholar
44.
O’Malley  D, Dewan  AA, Ohman-Strickland  PA, Gundersen  DA, Miller  SM, Hudson  SV.  Determinants of patient activation in a community sample of breast and prostate cancer survivors.   Psychooncology. 2018;27(1):132-140. doi:10.1002/pon.4387PubMedGoogle Scholar
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Views 970
    Citations 0
    Original Investigation
    November 4, 2020

    Association Between Patient Activation and Health Care Utilization After Thoracic and Abdominal Surgery

    Author Affiliations
    • 1Steinberg-Bernstein Centre for Minimally Invasive Surgery and Innovation, McGill University Health Centre, Montreal, Quebec, Canada
    • 2Department of Surgery, McGill University Health Centre, Montreal, Quebec, Canada
    • 3Division of Clinical Epidemiology and Biostatistics, McGill University, Montreal, Quebec, Canada
    JAMA Surg. 2021;156(1):e205002. doi:10.1001/jamasurg.2020.5002
    Key Points

    Question  Is patient activation (ie, knowledge, skills, motivation, confidence to participate in care) associated with unplanned health care utilization postdischarge after major surgery?

    Findings  In this cohort study of 653 patients, patients with low levels of activation had a higher risk of 30-day postdischarge unplanned health care utilization and complications and a longer length of hospital stay compared with patients with high levels of activation.

    Meaning  Patients at higher risk of costly unplanned health care utilization postdischarge can be identified prior to hospital discharge.

    Abstract

    Importance  Increased patient activation (PA) (ie, knowledge, skills, motivation, confidence to participate in care) may result in improved outcomes, especially in surgical settings.

    Objective  To estimate the extent to which PA is associated with 30-day postdischarge unplanned health care utilization after major thoracic or abdominal surgery.

    Design, Setting, and Participants  This cohort study was performed at 2 centers of a tertiary care hospital network between October 2017 and January 2019. Adult patients undergoing thoracic or abdominal surgery were included. Of 880 patients assessed for eligibility, 692 were deemed eligible, of whom 34 declined to participate, 1 withdrew consent, and 4 were excluded after consent.

    Exposures  Patient activation was measured immediately after surgery during the initial admission using the Patient Activation Measure (score range, 0-100). Patients were dichotomized into low and high PA groups using previously described thresholds (Patient Activation Measure score, ≤55.1).

    Main Outcomes and Measures  The primary outcome was unplanned 30-day postdischarge health care utilization (composite including emergency department and outpatient clinic visits and/or hospital readmission). Secondary outcomes were length of stay, 30-day emergency department visits, 30-day readmissions, and postoperative complications.

    Results  A total of 653 patients admitted for thoracic, general, colorectal, and gynecologic surgery were included in the study (mean [SD] age, 58 [15] years; 369 women [56%]; 366 [56%] had minimally invasive surgery; 52 [8%] had emergency surgery), of which 152 (23%) had a low level of PA. Baseline characteristics were similar between patients with low- and high-level PA. Low PA was associated with unplanned health care utilization (odds ratio [OR], 3.15; 95% CI, 2.05-4.86; P < .001), emergency department visits (OR, 1.64; 95% CI, 1.02-2.64; P = .04), complications (OR, 1.63; 95% CI, 1.11-2.41; P = .01), and length of stay (adjusted mean difference, 1.19 days; 95% CI, 0.06-2.33; P = .04). Low PA was not associated with a higher risk of readmission (adjusted OR, 1.04; 95% CI, 0.56-1.93; P = .90).

    Conclusions and Relevance  In this study, low level of PA was associated with postdischarge unplanned health care use, hospital stay, and complications after major surgery. Identification of patients with low activation may allow the implementation of interventions to improve health care knowledge and support self-management postdischarge.

    Introduction

    Unplanned health care utilization after hospital discharge is common, costly, and in certain cases, avoidable.1 Surgical patients account for a fifth of all postdischarge emergency department (ED) visits and a quarter of readmissions.2 Therefore, postdischarge hospital utilization is often used as a measure of health care quality.3 Strategies to improve the quality of surgery and resource utilization have tended to focus on clinician behavior and health care system organization.4 However, new interventions such as Enhanced Recovery Pathways and prehabilitation require patient participation to address health behaviors such as tobacco use, physical activity, and nutrition. Strategies to encourage patient and caregiver engagement are included in best practices guidelines for perioperative care, but the evidence supporting these interventions is limited.5

    Patient activation (PA) has emerged as an important pillar of a patient-centered model of care.6 Patient activation is a novel behavioral concept defined as the knowledge, skills, motivation, and confidence to participate in one’s own health care.7 Patient activation encompasses multiple key components of patient involvement, including self-efficacy and readiness to change health-related behaviors.7 Evidence supports that in patients with chronic medical conditions, a higher level of activation is associated with improved self-management behaviors, patient satisfaction, and health outcomes.8-12 Highly activated patients tend to have better problem-solving skills as well as peer support.10 Importantly, studies suggest that tailored interventions to increase PA may decrease unplanned health care use such as ED visits and readmissions13,14 and reduce health care costs.15

    Despite evidence supporting the role of PA in improving the outcomes of patients with chronic medical conditions,14 the effect of PA on postoperative outcomes remains unclear. If PA is associated with surgical outcomes, this would suggest a novel potentially modifiable target for quality of care improvement. The primary purpose of this study is to estimate the extent to which PA is associated with 30-day postdischarge unplanned health care visits (a composite including ED visits, outpatient clinic visits, and/or hospital readmission) after major thoracic or abdominal surgery. Secondarily, we explored the association of PA with hospital length of stay, ED visits, readmissions, and postoperative complications.

    Methods

    We conducted a prospective cohort study in 2 hospital sites of the McGill University Health Care Center (MUHC) between October 2017 and January 2019. The design and reporting of this study were in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.16

    Study Design

    This study was approved by the institutional ethics review board at MUHC. Patients who met inclusion criteria were approached the day after surgery by a member of the research team and provided written consent. Consenting patients were informed at time of enrollment of the primary outcome of the study and evaluated twice in the postoperative period. The hospital evaluation included the Patient Activation Measure, a socioeconomic questionnaire, a health status questionnaire, and review of the medical record for information about underlying diagnosis, comorbidities, and surgical procedure. At postdischarge at 30 days, patients were contacted by telephone by 1 researcher (T.D.) who was blinded to their baseline characteristics and activation level. Patients were asked to self-report any unplanned health care visits (ED, clinic or general practitioner visit, or hospital readmission) as well as complete the health status questionnaire. MUHC hospital records and the Dossier Santé Québec were reviewed for unplanned clinic visits or calls, ED visits, and readmissions to verify information given by patients. Although the Dossier Santé Québec does not capture health care visits per se, it contains all blood tests and imaging performed within the province’s public health network.

    Study Cohort

    Adult patients older than 18 years undergoing elective or emergency general, thoracic, colorectal, gynecologic, vascular, or urologic surgery were considered for inclusion. Patients were recruited from October 31, 2017, to April 6, 2018, and from October 1, 2018, to January 18, 2019. Trauma and transplant patients were excluded, as were patients requiring an intensive care unit stay of more than 3 days. Patients who did not speak or understand English or French or had neurologic or cognitive impairments that precluded them from answering questionnaires were excluded from the study.

    Study Measures

    Demographic data, including age, sex, body mass index, and comorbidities, were collected. Comorbidities were classified using the Charlson Comorbidity Index adjusted for age.17 The underlying diagnosis, type of surgical procedure, and surgical approach were recorded, as well as whether the procedure was elective or an emergency. Socioeconomic questionnaires included education level, employment status, type of work, and mean annual income. Patient activation was assessed using the Patient Activation Measure questionnaire supplied by Insignia Health on a research license.18 The survey includes 13 items evaluating knowledge, skills, beliefs, and confidence. An overall score (range, 0-100) categorized patients into 4 levels: level 1 (score, ≤47), level 2 (score, ≥47.1 to ≤55.1), level 3 (score, ≥55.2 to ≤72.4), and level 4 (score, ≥72.5). Consistent with other studies, patients were dichotomized into low PA (levels 1 and 2) and high PA (levels 3 and 4) groups.14 At the lowest level, patients were considered passive recipients of care, and at the highest level, patients were able to adopt new behaviors and maintain them under stress. The EuroQoL 5-dimension questionnaire was used to measure perceived health status19 including the EuroQoL 5-dimension 5-level assessing 5 dimensions (mobility, self-care, usual activity, pain/discomfort, and anxiety/depression; range, 1-5 with higher scores indicating greater impairment), and the EuroQoL 5-dimension visual analog scale assessing global health (range, 0-100, with 0 indicating worst health you can imagine and 100 indicating best health you can imagine). The minimal clinically important difference of the EuroQoL 5-dimension visual analog scale is estimated as 10 in most studies20 and a mean minimal clinically important difference of 0.074 has been reported for each of the 5 individual dimensions.21

    Study Outcomes

    The primary outcome variable of this study was the occurrence of any unplanned postdischarge health care visit 30 days after hospital discharge. This included hospital readmission, ED visits, and clinic visits (including general practitioner, surgical clinic, and nursing clinic). Visits to outside hospitals were identified by patient self-report and confirmed if a record in the Dossier Santé Québec was found. Hospital readmissions and ED visits were also analyzed separately as secondary outcomes because they are the costliest unplanned visits. Other secondary outcomes included index hospital length of stay (LOS), 30-day postoperative complications, return to work, and postdischarge health status. Postoperative complications were identified from the MUHC record and recorded up to 30 days after hospital discharge. Each complication was graded using the Clavien-Dindo classification22 and quantified using the Comprehensive Complication Index (range, 0-100).23,24

    Sample Size Calculation

    A previous study assessing the effect of PA on 30-day postdischarge hospital utilization in medical patients14 with 10% level 1 patients and 45% level 4 patients reported an incidence risk ratio of 1.75 (95% CI, 1.06-1.80; P < .001) of 30-day postdischarge hospitalization in level 1 compared with level 4 patients. Based on these data, accounting for 2-sided testing with an α of 0.05, power of 80%, and a 10% loss to follow-up, we estimated that a total of 650 patients would be required for this study.

    Statistical Analysis

    All patients enrolled in the study were included in the analysis. Summary descriptive statistics using frequency, proportion, mean (SD), or median (interquartile range [IQR]) were used to characterize the patient population. Demographics, patient characteristics, and rates of postoperative outcomes were compared between patients with high vs low level of PA using χ2 or Fisher exact test (categorical variables) and t test or 2-sided Mann-Whitney test (continuous variables). Missing data for clinical outcomes for patients who could not be reached at 30 days were handled using multiple imputations using chained equations (10 imputations). Using this method, missing items were estimated using a regression model from other observed data and repeated 10 times to generate 10 different imputed data sets. Uncertainty around the imputed point estimates incorporate the between (data sets) and within (variable) variances according to Rubin rules.25 Multiple logistic regression was used to determine the independent association of PA level on unplanned postdischarge health care utilization, ED visits, and readmissions adjusted for age, sex, education level, employment status, income, Charlson Comorbidity Index, surgical approach, and emergency surgery. Multiple logistic or linear regression was also used to determine the independent association of PA level with postoperative complications and LOS adjusted for age, sex, Charlson Comorbidity Index, surgical approach, and emergency surgery. For the analysis of LOS, we considered variables present at baseline and did not include postoperative complications. All analyses were conducted using Stata version 15 (StataCorp). Statistical significance was set at a 2-sided P value of .05.

    Results

    Of 1801 patients, 1261 underwent elective and 540 underwent emergency inpatient surgery in the specialties of interest during the study periods. Of these, 880 were assessed for eligibility, and 692 eligible patients were approached for recruitment, of whom 34 declined to participate, 1 withdrew consent voluntarily, and 4 patients were excluded because the surgical procedure did not meet inclusion criteria (Figure). Of the 653 patients included, the median (IQR) Patient Activation Measure score was 65.5 (55.6-75); 152 (23.3%) had a low level of PA (49 [7.5%] in level 1 and 103 [15.8%] in level 2) and 501 (76.7%) had a high level of PA (261 [40.0%] in level 3 and 240 [36.8%] in level 4). A total of 59 patients (9%) could not be reached for the telephone interview and were excluded from the analysis of 30-day patient-reported health state only. Losses to follow up were similar between the 2 groups (16 [11%] in the low PA and 43 [9%] in the high PA group).

    Baseline demographic and clinical characteristics of patients are reported in Table 1. Overall, 52 patients (8%) underwent emergency surgery. Patients with low activation were more likely to report being employed in a job requiring physical work. Other variables including comorbidity index, employment status, education, and income level were similar between groups. Patients with low PA reported significantly lower overall health (mean [95% CI], 52 [49-55] vs 59 [57-60]; P < .001) and significantly higher anxiety/depression (mean [95% CI], 1.90 [1.74-2.06] vs 1.63 [1.56-1.71]; P = .002) (Table 1).

    Unplanned health care utilization at 30-day postdischarge was significantly higher in patients with low PA compared with patients with high PA (64 [42%] vs 100 [20%]; P < .001). However, hospital readmissions were similar between the 2 groups (16 [11%] vs 55 [11%]) (Table 2). Of the ED visits, only 6 (5.7%) occurred at a site other than the MUHC and all readmissions were at our center only. Reasons for ED visits are reported in eTable 1 in the Supplement.

    Patients with low PA had longer initial hospital LOS compared with patients with high-level PA (median [IQR], 3.5 [2-6] vs 3 [1-5] days; P = .04). A total of 223 patients (34%) developed postoperative complications (Table 2). When assessing the timing of the complication, 94 of 223 patients (42%) with complications were diagnosed postdischarge, with a similar proportion in the 2 PA groups (30 [48%] vs 64 [40%]; P = .29). Global health state was higher 30 days after discharge compared with immediately postoperatively yet remained lower in patients with low compared with high PA (mean [95% CI], 70 [67-73] vs 77 [75-79]; P < .001). For employed patients reached at 30-day follow-up, those with low PA were less likely to have returned to work (18 [29%] vs 90 [45%]; P = .02).

    On multivariate logistic regression, low level of PA was associated with a higher risk of unplanned health care visits compared with high-level PA (adjusted odds ratio, 3.15; 95% CI, 2.05-4.86; P < .001; Table 3). Low-level PA was also associated with an increased risk of ED visits (adjusted odds ratio, 1.64; 95% CI, 1.02-2.64; P = .04) but was not associated with a higher risk of readmission (adjusted odds ratio, 1.04; 95% CI, 0.56-1.93; P = .90) (Table 3). Low activation was associated with an increased risk for complications (adjusted odds ratio, 1.63; 95% CI, 1.11-2.41; P = .01), as was Charlson Comorbidity Index, while minimally invasive surgery was associated with a decreased risk for complications (Table 4). Low-level PA was also associated with LOS (adjusted mean difference, 1.19 days; 95% CI, 0.06-2.33; P = .04), along with comorbidity index and emergency surgery (eTable 2 in the Supplement). The unadjusted univariate regressions are included in eTables 3 to 5 in the Supplement.

    Discussion

    Identifying modifiable risk factors for unplanned health care utilization is of significant importance not only from a health care system perspective, but also from patients’ perspective, when considering the distress associated with a hospital visit.26,27 In this study, a lower level of PA was associated with a higher risk of unplanned health care utilization 30 days after major thoracic and abdominal inpatient surgery. While low PA level was associated with increased risk of ED visits, it was not associated with hospital readmission. These results support that determining PA level preoperatively could help identify patients at higher risk of unplanned visits and prompt interventions to adequately prepare and support them after discharge.

    These findings are similar to previous literature supporting the association of PA with healthy behaviors and screening testing, clinical outcomes, and hospital and ED visits, independent of sociodemographic characteristics.9,10,14,28-30 However, most of this work was in the context of chronic medical conditions. To our knowledge, this is the first study of the role of PA in patients undergoing thoracic and abdominal surgery. A few previous studies assessed the role of PA in the functional recovery of patients undergoing orthopedic surgery.31-33 Similar to previous studies, we found that PA was independent of age, sex, level of education, employment status, and annual income and therefore may be a novel and potentially modifiable variable.

    Patients with lower levels of activation were at higher risk for developing a complication. The Clavien complication severity grade distribution did not differ between the 2 groups, but patients in the low PA group had a higher median comprehensive complication index, which is considered a more sensitive indicator of the total burden of complications.23 Patient activation has been shown to correlate with patient participation in rehabilitation programs after orthopedic surgery, which may explain the relationship to postoperative complications.31,34 We considered whether the higher risk of complications was the reason for the unplanned visit or was rather a reporting bias in that patients with lower levels of PA were also more likely to seek medical attention postdischarge and be evaluated. However, the proportion of patients presenting with a postdischarge complication was similar between the PA levels, suggesting the occurrence of complications was not the sole driver of the unplanned visit.

    While low PA was associated with a higher risk of ED visits, these did not result in a higher risk of readmissions, in contrast to patients with chronic medical conditions.14 This suggests that the reason for the ED visit could have potentially been managed in a different setting.1,35-38 Having the ability to identify patients at risk for potentially avoidable ED visits may help build a more tailored and patient-centered discharge plan.3,4,39,40 The answers provided on the Patient Activation Measure questionnaire may suggest specific areas where patients need help, such as further education, skills development, or nursing or social worker support. Importantly, PA level is changeable and can be increased through tailored interventions focused on skills training and encouragement of a sense of ownership of health.41 Furthermore, in chronic diseases, activation-focused interventions including coaching, education, and peer support can result in sustained improvements in self-management behaviors and clinical outcomes, as well as reduced use of health care services.11,12,37

    Surgical patients differ from chronic medical patients in the acute nature of their therapeutic management.42 Major surgery results in a predictable decline in functional status and health-related quality of life, which requires a period of recovery of weeks to months. In this study, patients with a low level of PA reported lower overall health state and more impairment in mobility, ability to perform usual activity, pain/discomfort, and anxiety/depression compared with patients with a high level of PA. Moreover, for employed patients, a higher proportion of patients with high PA levels returned to work within 30 days after discharge. This may be because patients with low PA were more likely to have a physical job and reported more impairment in the domains of pain/discomfort, activities, mobility, and anxiety/depression.

    Strengths and Limitations

    This study has several strengths. It included a relatively large sample of patients undergoing a wide range of major surgical procedures, adding to its generalizability. The review of the provincial health record identified ED visit testing occurring at outside centers. Patient-reported unplanned health care use allowed for the capture of office and clinic visits that are not included in hospital and provincial records. Patients were informed of the primary outcome at time of enrollment, which sensitized documentation of unplanned health care uses. Observer bias was also minimized because the researcher assessing outcome variables was unaware of the patient baseline characteristics and activation level.

    The results of this study should be interpreted in light of several limitations. First, we could not reach 59 patients 1 month after surgery. However, the losses to follow-up were similar between the 2 groups (11% in the low PA group and 9% in the high PA group). We relied on patients to report clinic and general practitioner visits that were not recorded in the provincial record and did not have access to clinical records from outside the MUHC. In addition, the main finding of the association between PA and unplanned visits was maintained when only considering ED visits, which were fully captured, with only 6% occurring at an outside institution. Second, although this is a relatively large study, we approached less than half of the patients undergoing potentially relevant procedures during the study period. This may be because of delayed recording of surgical admissions in the operating room database we used to identify patients. In addition, while the study protocol specified inclusion of both emergency and elective patients, our patient identification strategy resulted in the unexpected underrepresentation of emergency surgery patients (8% of the study population vs 30% of procedures during the study periods). Although we adjusted for this variable in our regression models, this may limit the applicability of the results to elective procedures only and the effect of PA in emergency surgery warrants future investigation. Third, this study was performed at an academic medical center in Canada, and results may only be generalizable to similar settings. However, the distribution of the PA levels was consistent with previous studies from other geographic locations and patient settings.13,14,43,44

    Conclusions

    The results of this study suggest that patients with low levels of activation are at increased risk of early unplanned health care utilization postdischarge after major surgery, including ED visits, without an increased risk in readmission. Lower-level PA was also associated with increased risk of 30-day complications, longer hospital stay, and lower health-related quality of life during the recovery period. As a more patient-centered approach to surgical care is increasingly advocated, the identification of patients with lower levels of activation could prompt the provision of additional support through targeted education or other resources to improve postoperative outcomes.

    Back to top
    Article Information

    Corresponding Author: Liane S. Feldman, MD, McGill University Health Centre, 1650 Cedar Ave, D6-156, Montreal, QC H3G 1A4, Canada (liane.feldman@mcgill.ca).

    Accepted for Publication: August 12, 2020.

    Published Online: November 4, 2020. doi:10.1001/jamasurg.2020.5002

    Author Contributions: Drs Feldman and Dumitra had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

    Concept and design: Dumitra, Fiore, Mayo, Lee, Feldman.

    Acquisition, analysis, or interpretation of data: Dumitra, Ganescu, Hu, Fiore, Kaneva, Lee, Liberman, Chaudhury, Ferri, Feldman.

    Drafting of the manuscript: Dumitra, Feldman.

    Critical revision of the manuscript for important intellectual content: All authors.

    Statistical analysis: Dumitra, Fiore, Mayo, Lee.

    Obtained funding: Dumitra.

    Administrative, technical, or material support: Dumitra, Ganescu, Hu, Fiore, Kaneva, Liberman.

    Supervision: Kaneva, Mayo, Lee, Liberman, Chaudhury, Ferri, Feldman.

    Conflict of Interest Disclosures: Dr Fiore reports grants from Merck and personal fees from Shionogi outside the submitted work. Dr Lee reports grants from Johnson & Johnson outside the submitted work. Dr Liberman reports personal fees from Merck, Servier, and Ipsen outside the submitted work. No other disclosures were reported.

    Funding/Support: Dr Dumitra received a salary award from Fonds de Recherche du Québec - Santé (FRQS).

    Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

    References
    1.
    Hsia  RY, Niedzwiecki  M.  Avoidable emergency department visits: a starting point.   Int J Qual Health Care. 2017;29(5):642-645. doi:10.1093/intqhc/mzx081PubMedGoogle Scholar
    2.
    All-Cause Readmission to Acute Care and Return to the Emergency Department. Canadian Institute for Health Information; 2012. https://secure.cihi.ca/free_products/Readmission_to_acutecare_en.pdf
    3.
    Thanh  NX, Chuck  AW, Wasylak  T,  et al.  An economic evaluation of the Enhanced Recovery After Surgery (ERAS) multisite implementation program for colorectal surgery in Alberta.   Can J Surg. 2016;59(6):415-421. doi:10.1503/cjs.006716PubMedGoogle Scholar
    4.
    Gillis  C, Gill  M, Marlett  N,  et al.  Patients as partners in Enhanced Recovery After Surgery: a qualitative patient-led study.   BMJ Open. 2017;7(6):e017002. doi:10.1136/bmjopen-2017-017002PubMedGoogle Scholar
    5.
    Fearon  KC, Ljungqvist  O, Von Meyenfeldt  M,  et al.  Enhanced recovery after surgery: a consensus review of clinical care for patients undergoing colonic resection.   Clin Nutr. 2005;24(3):466-477. doi:10.1016/j.clnu.2005.02.002PubMedGoogle Scholar
    6.
    Glasgow  RE, Orleans  CT, Wagner  EH.  Does the chronic care model serve also as a template for improving prevention?   Milbank Q. 2001;79(4):579-612, iv-v. iv-v. doi:10.1111/1468-0009.00222PubMedGoogle Scholar
    7.
    Hibbard  JH, Stockard  J, Mahoney  ER, Tusler  M.  Development of the Patient Activation Measure (PAM): conceptualizing and measuring activation in patients and consumers.   Health Serv Res. 2004;39(4 Pt 1):1005-1026. doi:10.1111/j.1475-6773.2004.00269.xPubMedGoogle Scholar
    8.
    Hibbard  JH, Mahoney  ER, Stock  R, Tusler  M.  Do increases in patient activation result in improved self-management behaviors?   Health Serv Res. 2007;42(4):1443-1463. doi:10.1111/j.1475-6773.2006.00669.xPubMedGoogle Scholar
    9.
    Remmers  C, Hibbard  J, Mosen  DM, Wagenfield  M, Hoye  RE, Jones  C.  Is patient activation associated with future health outcomes and healthcare utilization among patients with diabetes?   J Ambul Care Manage. 2009;32(4):320-327. doi:10.1097/JAC.0b013e3181ba6e77PubMedGoogle Scholar
    10.
    Greene  J, Hibbard  JH.  Why does patient activation matter? an examination of the relationships between patient activation and health-related outcomes.   J Gen Intern Med. 2012;27(5):520-526. doi:10.1007/s11606-011-1931-2PubMedGoogle Scholar
    11.
    Hibbard  JH, Greene  J, Shi  Y, Mittler  J, Scanlon  D.  Taking the long view: how well do patient activation scores predict outcomes four years later?   Med Care Res Rev. 2015;72(3):324-337. doi:10.1177/1077558715573871PubMedGoogle Scholar
    12.
    Greene  J, Hibbard  JH, Alvarez  C, Overton  V.  Supporting patient behavior change: approaches used by primary care clinicians whose patients have an increase in activation levels.   Ann Fam Med. 2016;14(2):148-154. doi:10.1370/afm.1904PubMedGoogle Scholar
    13.
    Hibbard  JH, Greene  J, Tusler  M.  Improving the outcomes of disease management by tailoring care to the patient’s level of activation.   Am J Manag Care. 2009;15(6):353-360.PubMedGoogle Scholar
    14.
    Mitchell  SE, Gardiner  PM, Sadikova  E,  et al.  Patient activation and 30-day post-discharge hospital utilization.   J Gen Intern Med. 2014;29(2):349-355. doi:10.1007/s11606-013-2647-2PubMedGoogle Scholar
    15.
    Harvey  L, Fowles  JB, Xi  M, Terry  P.  When activation changes, what else changes? the relationship between change in Patient Activation Measure (PAM) and employees’ health status and health behaviors.   Patient Educ Couns. 2012;88(2):338-343. doi:10.1016/j.pec.2012.02.005PubMedGoogle Scholar
    16.
    Vandenbroucke  JP, von Elm  E, Altman  DG,  et al; STROBE Initiative.  Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): explanation and elaboration.   Int J Surg. 2014;12(12):1500-1524. doi:10.1016/j.ijsu.2014.07.014PubMedGoogle Scholar
    17.
    Charlson  M, Szatrowski  TP, Peterson  J, Gold  J.  Validation of a combined comorbidity index.   J Clin Epidemiol. 1994;47(11):1245-1251. doi:10.1016/0895-4356(94)90129-5PubMedGoogle Scholar
    18.
    Hibbard  JH, Mahoney  ER, Stockard  J, Tusler  M.  Development and testing of a short form of the Patient Activation Measure.   Health Serv Res. 2005;40(6 Pt 1):1918-1930. doi:10.1111/j.1475-6773.2005.00438.xPubMedGoogle Scholar
    19.
    Devlin  NJ, Krabbe  PF.  The development of new research methods for the valuation of EQ-5D-5L.   Eur J Health Econ. 2013;14(suppl 1):S1-S3. doi:10.1007/s10198-013-0502-3PubMedGoogle Scholar
    20.
    Myles  PS, Myles  DB, Galagher  W,  et al.  Measuring acute postoperative pain using the visual analog scale: the minimal clinically important difference and patient acceptable symptom state.   Br J Anaesth. 2017;118(3):424-429. doi:10.1093/bja/aew466PubMedGoogle Scholar
    21.
    Walters  SJ, Brazier  JE.  Comparison of the minimally important difference for two health state utility measures: EQ-5D and SF-6D.   Qual Life Res. 2005;14(6):1523-1532. doi:10.1007/s11136-004-7713-0PubMedGoogle Scholar
    22.
    Clavien  PA, Barkun  J, de Oliveira  ML,  et al.  The Clavien-Dindo classification of surgical complications: five-year experience.   Ann Surg. 2009;250(2):187-196. doi:10.1097/SLA.0b013e3181b13ca2PubMedGoogle Scholar
    23.
    Slankamenac  K, Nederlof  N, Pessaux  P,  et al.  The comprehensive complication index: a novel and more sensitive endpoint for assessing outcome and reducing sample size in randomized controlled trials.   Ann Surg. 2014;260(5):757-762. doi:10.1097/SLA.0000000000000948PubMedGoogle Scholar
    24.
    Dumitra  TC, Trepanier  M, Fiore  JF  Jr,  et al.  The relationship of two postoperative complication grading schemas with postoperative quality of life after elective colorectal surgery.   Surgery. 2019;166(4):663-669. doi:10.1016/j.surg.2019.05.058PubMedGoogle Scholar
    25.
    Rubin  DB, Schenker  N.  Multiple imputation in health-care databases: an overview and some applications.   Stat Med. 1991;10(4):585-598. doi:10.1002/sim.4780100410PubMedGoogle Scholar
    26.
    Lateef  F.  Patient expectations and the paradigm shift of care in emergency medicine.   J Emerg Trauma Shock. 2011;4(2):163-167. doi:10.4103/0974-2700.82199PubMedGoogle Scholar
    27.
    Sabbatini  AK, Kocher  KE, Basu  A, Hsia  RY.  In-hospital outcomes and costs among patients hospitalized during a return visit to the emergency department.   JAMA. 2016;315(7):663-671. doi:10.1001/jama.2016.0649PubMedGoogle Scholar
    28.
    Greene  J, Hibbard  JH, Sacks  R, Overton  V, Parrotta  CD.  When patient activation levels change, health outcomes and costs change, too.   Health Aff (Millwood). 2015;34(3):431-437. doi:10.1377/hlthaff.2014.0452PubMedGoogle Scholar
    29.
    Parchman  ML, Zeber  JE, Palmer  RF.  Participatory decision making, patient activation, medication adherence, and intermediate clinical outcomes in type 2 diabetes: a STARNet study.   Ann Fam Med. 2010;8(5):410-417. doi:10.1370/afm.1161PubMedGoogle Scholar
    30.
    Schmaderer  MS, Zimmerman  L, Hertzog  M, Pozehl  B, Paulman  A.  Correlates of patient activation and acute care utilization among multimorbid patients.   West J Nurs Res. 2016;38(10):1335-1353. doi:10.1177/0193945916651264PubMedGoogle Scholar
    31.
    Skolasky  RL, Mackenzie  EJ, Wegener  ST, Riley  LH  III.  Patient activation and adherence to physical therapy in persons undergoing spine surgery.   Spine (Phila Pa 1976). 2008;33(21):E784-E791. doi:10.1097/BRS.0b013e31818027f1PubMedGoogle Scholar
    32.
    Skolasky  RL, Mackenzie  EJ, Riley  LH  III, Wegener  ST.  Psychometric properties of the Patient Activation Measure among individuals presenting for elective lumbar spine surgery.   Qual Life Res. 2009;18(10):1357-1366. doi:10.1007/s11136-009-9549-0PubMedGoogle Scholar
    33.
    Skolasky  RL, Mackenzie  EJ, Wegener  ST, Riley  LH.  Patient activation and functional recovery in persons undergoing spine surgery.   Orthopedics. 2011;34(11):888. doi:10.3928/01477447-20110922-04PubMedGoogle Scholar
    34.
    Andrawis  J, Akhavan  S, Chan  V, Lehil  M, Pong  D, Bozic  KJ.  Higher preoperative patient activation associated with better patient-reported outcomes after total joint arthroplasty.   Clin Orthop Relat Res. 2015;473(8):2688-2697. doi:10.1007/s11999-015-4247-4PubMedGoogle Scholar
    35.
    Brennan  JJ, Chan  TC, Killeen  JP, Castillo  EM.  Inpatient readmissions and emergency department visits within 30 days of a hospital admission.   West J Emerg Med. 2015;16(7):1025-1029. doi:10.5811/westjem.2015.8.26157PubMedGoogle Scholar
    36.
    Sources of Potentially Avoidable Emergency Department Visits. Canadian Institute for Health Information; 2014. https://secure.cihi.ca/free_products/ED_Report_ForWeb_EN_Final.pdf https://secure.cihi.ca/free_products/ED_Report_ForWeb_EN_Final.pdf2014
    37.
    Hibbard  JH, Greene  J, Sacks  RM, Overton  V, Parrotta  C.  Improving population health management strategies: identifying patients who are more likely to be users of avoidable costly care and those more likely to develop a new chronic disease.   Health Serv Res. 2017;52(4):1297-1309. doi:10.1111/1475-6773.12545PubMedGoogle Scholar
    38.
    Lash  RS, Bell  JF, Reed  SC,  et al.  A systematic review of emergency department use among cancer patients.   Cancer Nurs. 2017;40(2):135-144. doi:10.1097/NCC.0000000000000360PubMedGoogle Scholar
    39.
    Gustafsson  UO, Scott  MJ, Schwenk  W,  et al; Enhanced Recovery After Surgery (ERAS) Society, for Perioperative Care; European Society for Clinical Nutrition and Metabolism (ESPEN); International Association for Surgical Metabolism and Nutrition (IASMEN).  Guidelines for perioperative care in elective colonic surgery: Enhanced Recovery After Surgery (ERAS) Society recommendations.   World J Surg. 2013;37(2):259-284. doi:10.1007/s00268-012-1772-0PubMedGoogle Scholar
    40.
    Wood  T, Aarts  MA, Okrainec  A,  et al; iERAS group.  Emergency room visits and readmissions following Implementation of an Enhanced Recovery After Surgery (iERAS) program.   J Gastrointest Surg. 2018;22(2):259-266. doi:10.1007/s11605-017-3555-2PubMedGoogle Scholar
    41.
    Gleason-Comstock  J, Streater  A, Ager  J,  et al.  Patient education and follow-up as an intervention for hypertensive patients discharged from an emergency department: a randomized control trial study protocol.   BMC Emerg Med. 2015;15:38. doi:10.1186/s12873-015-0052-3PubMedGoogle Scholar
    42.
    Yun  PS, MacDonald  CL, Orne  J,  et al.  A Novel surgical patient engagement model: a qualitative study of postoperative patients.   J Surg Res. 2020;248:82-89. doi:10.1016/j.jss.2019.11.025PubMedGoogle Scholar
    43.
    Blakemore  A, Hann  M, Howells  K,  et al.  Patient activation in older people with long-term conditions and multimorbidity: correlates and change in a cohort study in the United Kingdom.   BMC Health Serv Res. 2016;16(1):582. doi:10.1186/s12913-016-1843-2PubMedGoogle Scholar
    44.
    O’Malley  D, Dewan  AA, Ohman-Strickland  PA, Gundersen  DA, Miller  SM, Hudson  SV.  Determinants of patient activation in a community sample of breast and prostate cancer survivors.   Psychooncology. 2018;27(1):132-140. doi:10.1002/pon.4387PubMedGoogle Scholar
    ×