Research ArticleOriginal ResearchA

A Cluster Randomized Trial of Primary Care Practice Redesign to Integrate Behavioral Health for Those Who Need It Most: Patients With Multiple Chronic Conditions

Benjamin Littenberg, Jessica Clifton, Abigail M. Crocker, Laura-Mae Baldwin, Levi N. Bonnell, Ryan E. Breshears, Peter Callas, Prama Chakravarti, Kelly Clark/Keefe, Deborah J. Cohen, Frank V. deGruy, Lauren Eidt-Pearson, William Elder, Chester Fox, Sylvie Frisbie, Katie Hekman, Juvena Hitt, Jennifer Jewiss, David C. Kaelber, Kairn Stetler Kelley, Rodger Kessler, Jennifer B. O’Rourke-Lavoie, George S. Leibowitz, C. R. Macchi, Matthew P. Martin, Mark McGovern, Brenda Mollis, Daniel Mullin, Zsolt Nagykaldi, Lisa W. Natkin, Wilson Pace, Richard G. Pinckney, Douglas Pomeroy, Paula Reynolds, Gail L. Rose, Sarah Hudson Scholle, William J. Sieber, Jeni Soucie, Terry Stancin, Kurt C. Stange, Kari A. Stephens, Kathryn Teng, Elizabeth Needham Waddell and Constance van Eeghen
The Annals of Family Medicine November 2023, 21 (6) 483-495; DOI: https://doi.org/10.1370/afm.3027

Article Figures & Data

Figures

  • Figure 1.
    Figure 1.

    CONSORT participant flow diagram.

    CONSORT = Consolidated Standards of Reporting Trials.

    a Eligible but chose not to participate.

    b Wanted to participate but not eligible.

    c Eligible but volunteered to pilot the intervention before study start.

  • Figure 2.
    Figure 2.
    Figure 2.

    Effect of active group assignment on patient and practice outcomes.

    PIP = Practice Integration Profile; PROMIS-29 = Patient-Reported Outcomes Measurement Information System.

    Note: The center point of each bar is the mean adjusted effect (regression coefficient) of the intervention on a specific outcome, with the bars representing 95% CI. Values >0 indicate that the intervention was associated with an increase in the outcome measured. Bars that do not cross 0 indicate statistical significance (P < .05).

    a Greater score indicates improved outcome.

  • Figure 3.
    Figure 3.
    Figure 3.

    Association of practice integration and patient outcomes.

    PIP = Practice Integration Profile.

    Note: The center point of each bar is the mean adjusted association (regression coefficient) of the baseline median total PIP score with specific patient outcomes, adjusting for potential confounders and using a random intercept for each practice. The bars represent 95% CI. Values >0 indicate that the intervention was associated with an increase in the outcome measured. Bars that do not cross 0 indicate statistical significance (P < .05).

    a Greater score indicates improved outcome.

  • Figure 4.
    Figure 4.
    Figure 4.

    Association of intervention completion in active group (n = 20) and change in practice integration and patient-reported outcomes.

    PIP = Practice Integration Profile; PROMIS-29 = Patient-Reported Outcomes Measurement Information System.

    Note: The center point of each bar is the mean adjusted effect (regression coefficient) of the completion of the intervention on a specific outcome, with the bars representing 95% CI. Values >0 indicate that completion of the intervention was associated with an increase in the outcome measured. Bars that do not cross 0 indicate statistical significance (P < .05).

    a Greater score indicates improved outcome.

Tables

  • Table 1.

    Patient Characteristics at Baseline

    TotalControlActiveP Value
    No. (%)Mean (SD)No. (%)Mean (SD)No. (%)Mean (SD)
    na2,426  1,459   967
    Age,b y61.9 (13.1)61.9 (13.4)61.9 (12.7).49
    Sexb.15
        Female1,588 (65.5)   970 (66.5)   618 (63.9)
        Male   832 (34.3)   483 (33.1)   349 (36.1)
    Raceb.01
        White1,859 (76.8)1,157 (79.3)   702 (72.9)
        Black or African American   281 (11.6)   154 (10.6)   127 (13.2)
        Asian     71 (2.9)     35 (2.4)     36 (3.7)
        Native Hawaiian/Other Pacific Islander     37 (1.5)     17 (1.2)     20 (2.1)
        American Indian or Alaskan Native     21 (0.9)     13 (0.9)       8 (0.8)
        Other   153 (6.3)     83 (5.7)     70 (7.3)
    Ethnicityb.14
        Non-Hispanic2,207 (91.5)1,342 (92.4)   865 (90.1)
        Hispanic   180 (7.5)     96 (6.6)     84 (8.8)
        Prefer not to say     26 (1.1)     15 (1.0)     11 (1.1)
    Marital statusb.13
        Married1,082 (44.8)   621 (42.7)   461 (47.9)
        Divorced   517 (21.4)   318 (21.9)   199 (20.7)
        Never married   405 (16.8)   249 (17.1)   156 (16.2)
        Widowed   290 (12.0)   183 (12.6)   107 (11.1)
        Separated     62 (2.6)     44 (3.0)     18 (1.9)
        Living as married     61 (2.5)     39 (2.7)     22 (2.3)
    Employmentb.11
        Retired   864 (36.6)   525 (36.8)   339 (36.2)
        Disabled   627 (26.5)   372 (26.1)   255 (27.2)
        Full-time   486 (20.6)   274 (19.2)   212 (22.6)
        Part-time   205 (8.7)   136 (9.5)     69 (7.4)
        Homemaker     94 (4.0)     60 (4.2)     34 (3.6)
        Unemployed/looking     68 (2.9)     43 (3.0)     25 (2.7)
        Student     15 (0.6)     13 (0.9)       2 (0.2)
        Other       3 (0.1)       2 (0.1)       1 (0.1)
    Annual household income,b $
        <15,000   681 (29.5)   416 (30.0)   265 (28.7).04
        15,000-29,999   509 (22.0)   308 (22.2)   201 (21.8)
        30,000-44,999   281 (12.2)   180 (13.0)   101 (10.9)
        45,000-59,999   186 (8.0)   116 (8.4)     70 (7.6)
        60,000-74,999   190 (8.2)   121 (8.7)     69 (7.5)
        75,000-99,999   188 (8.1)   103 (7.4)     85 (9.2)
        >100,000   277 (12.0)   144 (10.4)   133 (14.4)
    Educationb.92
        <Grade 9     56 (2.4)     32 (2.2)     24 (2.5)
        Grades 9-12, no diploma   208 (8.8)   123 (8.6)     85 (9.0)
        High school graduate (including GED)1,004 (42.2)   608 (42.6)   396 (41.7)
        Associate degree   378 (15.9)   222 (15.5)   156 (16.4)
        Bachelor’s degree   387 (16.3)   229 (16.0)   158 (16.6)
        Graduate or professional degree   344 (14.5)   214 (15.0)   130 (13.7)
    Chronic conditionsb
        Arthritis1,023 (42.2)   613 (42.0)   410 (42.4).85
        Asthma   545 (22.5)   316 (21.7)   229 (23.7).24
        Chronic obstructive pulmonary disease   330 (13.6)   214 (14.7)   116 (12.0).06
        Chronic pain2,037 (84.0)1,213 (83.1)   824 (85.2).17
        Diabetes (nongestational)1,075 (44.3)   671 (46.0)   404 (41.8).04
        Heart failure   188 (7.8)   119 (8.2)     69 (7.1).36
        Hypertension2,012 (82.9)1,207 (82.7)   805 (83.3).74
        Irritable bowel syndrome   102 (4.2)     56 (3.8)     46 (4.8).27
        Anxiety   830 (34.2)   491 (33.7)   339 (35.1).48
        Depression1,136 (46.8)   697 (47.8)   439 (45.4).25
        Insomnia   570 (23.5)   339 (23.2)   231 (23.9).71
        Any substance use disorder   559 (23.0)   339 (23.2)   220 (22.8).78
        Tobacco use   445 (18.3)   274 (18.8)   171 (17.7).49
        Alcohol use disorder   155 (6.4)     93 (6.4)     62 (6.4).97
    Mean number of chronic conditionsb4.4 (1.6)4.4 (1.6)4.3 (1.6).6
    Neighborhood characteristicsb (home census tract)
        Social Deprivation Index (higher score indicates more deprivation)52.6 (27.7)53.5 (26.9)51.2 (29.0).05
        Rural   477 (19.8)   341 (23.4)   136 (14.3)<.001
        Population density, persons per square mile  3,768
    (6,656)    		  2,978
    (3,315)    		  4,980
    (9,645)    		<.001
    Primary outcomes – PROMIS-29 t-scoresc
        Anxiety53.9 (10.0)53.8 (9.9)54.1 (10.1).34
        Depression52.7 (9.8)52.7 (9.8)52.9 (9.7).46
        Fatigue52.6 (10.4)52.6 (10.4)52.5 (10.4).66
        Sleep disturbance53.1 (8.9)52.9 (8.8)53.3 (8.9).26
        Pain interference58.1 (10.0)58.0 (10.1)58.2 (10.0).54
        Pain intensity  4.5 (2.8)  4.4 (2.7)  4.5 (2.8).58
        Social participationc48.3 (9.9)48.4 (9.8)48.1 (10.0).65
        Physical functionc43.6 (9.4)43.5 (9.4)43.7 (9.3).49
    Secondary outcomes
        PROMIS-29 Physical Health Summary t-scorec45.9 (9.4)45.8 (9.5)46.0 (9.3).7
        PROMIS-29 Mental Health Summary t-scorec50.3 (8.8)50.4 (8.8)50.2 (8.9).67
        CARE total scorec  4.3 (0.9)  4.3 (0.9)  4.3 (0.9).03
        MMAS total scorec  1.0 (1.1)  1.0 (1.1)  1.0 (1.1).57
        ED visits in past year  1.0 (1.8)  1.0 (1.7)  1.1 (1.8).94
        Health care visits in past month  2.5 (2.7)  2.4 (2.6)  2.6 (2.8).05
        Hospital days in past year  1.0 (2.4)  1.0 (2.3)  1.0 (2.5).28
        Restricted activity days  1.1 (2.5)  1.1 (2.5)  1.1 (2.5).86
        Metabolic equivalents (DASI)c  6.4 (2.0)  6.4 (1.9)  6.4 (2.0).96
        PHQ-9 total score  6.4 (6.1)  6.3 (6.1)  6.5 (6.1).48
        GAD-7 total score  4.5 (5.2)  4.4 (5.2)  4.7 (5.3).17
        Asthma Symptom Utility Index  0.8 (0.2)  0.8 (0.2)  0.8 (0.2).57
        GAIN Lifetime SUD screener  1.7 (1.9)  1.7 (1.9)  1.8 (1.9).55
        GAIN 1 Year SUD screener  0.6 (1.2)  0.6 (1.1)  0.6 (1.2).57
        GAIN 3 Month SUD screener  0.5 (1.0)  0.5 (1.0)  0.5 (1.0).62
        GAIN 1 Month SUD screener  0.5 (0.9)  0.4 (0.9)  0.5 (1.0).63
        Unhealthy alcohol use   171 (9.6)   109 (9.7)     62 (9.6).94
        Patient-Centeredness Indexc83.9 (15.9)  84.0 (15.9)83.8 (15.7).62

    • CARE = Consultation and Relational Empathy; DASI = Duke Activity Status Index; ED = emergency department; GAD-7 = 7-item Generalized Anxiety Disorder screener; GAIN = Global Appraisal of Individual Needs; GED = General Education Development; MMAS = Morisky Medication Adherence Scale; PHQ-9 = 9-item Patient Health Questionnaire; PROMIS-29 = Patient-Reported Outcomes Measurement Information System; SUD = substance use disorder.

    • Note: Wilcoxon rank-sum test used for continuous variables; χ2 test used for categorical variables.

    • ↵a No. varies for some characteristics.

    • ↵b Included as possible confounding effect. Other possible confounders included various county characteristics (age, racial composition, ethnicity, sex, income, education, employment, population density, urban/rural status).

    • ↵c Greater score indicates better health. T-score is a measure in which 50 is the mean of a relevant population, and 10 is the SD of that population.

  • Table 2.

    Practice Characteristics at Baseline

    Total n = 42Control n = 22Active n = 20P Value
    No. (%)Mean (SD)No. (%)Mean (SD)No. (%)Mean (SD)
    Practice specialtya  .92
        Internal medicine  7 (17)  4 (18)  3 (15)
        Family medicine20 (48)11 (50)  9 (45)
        Mixed15 (36)  7 (32)  8 (40)
    Organization typea,b
        Community health center15 (36)  7 (32)  8 (40)  .75
        Hospital20 (48)10 (45)10 (50)>.99
        Private  4 (10)  3 (14)  1 (5)  .61
        Academic19 (45)  9 (41)10 (50)  .76
    Number of patient visits per yeara27,03930,10423,668  .51
    (19,536)(24,210)(12,363)
    Baseline primary care provider FTEa  6.0 (3.2)  6.1 (3.6)  5.9 (2.7)  .96
    Baseline BH provider FTEa  1.5 (1.1)  1.3 (0.7)  1.7 (1.4)  .27
    Patients cared for by the practice per yeara    9,285
      (5,066)    		    9,419
      (5,599)    		    9,138
      (4,549)    		  .94
    Resident training sitea16 (38)  7 (32)  9 (45)  .53
    Nonprofita37 (88)18 (82)19 (95)  .35
    Tenure of on-site BH services, y  6.3 (6.4)  6.6 (7.4)  5.9 (5.3)  .86
    Geographic region.9
        Pacific Northwest  3 (7)  2 (9)  1 (5)
        Mountain  8 (19)  4 (18)  4 (20)
        South  8 (19)  4 (18)  4 (20)
        New England  9 (21)  6 (27)  3 (15)
        Mid-Atlantic & Great Lakes  6 (14)  3 (14)  3 (15)
        West Coast & Hawaii  8 (19)  3 (14)  5 (25)
    Urban by RUCA35 (83)17 (77)18 (90)  .41
    County social deprivation index45 (22)44 (21)46 (23)  .76
    PIP baseline total mediana59 (14)59 (17)59 (11)  .96

    • BH = behavioral health; FTE = full-time equivalent employee; PIP = Practice Integration Profile; RUCA = rural-urban commuting area.

    • Note: P values calculated by Wilcoxon rank-sum test or Fisher exact test.

    • ↵a Included as possible confounding effect.

    • ↵b Multiple types were possible; percentages may not sum to 100%.

  • Table 3.

    Adjusted Effects of Assignment to Active Group on Outcomes

    OutcomeNo.aEffect95% CIP Value
    Primary outcomes
        Anxiety2,426  0.08−0.53 to 0.69.79
        Depression2,426  0.21−0.37 to 0.79.48
        Fatigue2,426  0.07−0.54 to 0.68.83
        Sleep disturbance2,426−0.05−0.58 to 0.49.86
        Pain interference2,426  0.19−0.40 to 0.79.52
        Pain intensity2,426  0.1−0.08 to 0.28.27
        Social participationb2,426−0.06−0.71 to 0.59.86
        Physical functionb2,426  0.1−0.39 to 0.60.69
    Secondary outcomes
        Physical health summaryb2,426  0.09−0.39 to 0.56.72
        Mental health summaryb2,426−0.06−0.51 to 0.39.79
        CARE total scoreb2,383−0.03−0.11 to 0.05.49
        MMAS total scoreb2,310−0.03−0.11 to 0.05.46
        ED visits in past year2,369−0.01−0.14 to 0.11.82
        Health care visits in past month2,363−0.02−0.29 to 0.26.9
        Hospital days in past year2,363−0.04−0.29 to 0.21.78
        Utilization category2,376  0.07−0.13 to 0.27.48
        Restricted activity days2,301  0.16−0.02 to 0.35.08
        Metabolic equivalents (DASI)b2,127  0.03−0.09 to 0.15.64
        PHQ-9 total score2,290  0.2−0.13 to 0.54.24
        PHQ category2,290  0.14−0.04 to 0.32.12
        GAD-7 total score2,317  0.03−0.27 to 0.32.86
        GAD category2,317−0.06−0.25 to 0.13.53
        Asthma symptom utility index  440−0.02−0.05 to 0.01.17
        GAIN Lifetime SUD screener  376−0.06−0.35 to 0.24.71
        GAIN 1-Year SUD screener  376−0.15−0.37 to 0.06.16
        GAIN 3-Month SUD screener  376  0−0.19 to 0.18.96
        GAIN 1-Month SUD screener  376−0.04−0.22 to 0.14.65
        Alcohol use category1,747−0.03−0.36 to 0.29.84
        Patient-centeredness indexb2,261−0.8−2.3 to 0.7.29
    Practice Integration Profile domains
        Totalb    423.8−3.6 to 11.2.3
        Workflowb    429.31.7 to 16.9.02
        Clinical servicesb    422.6−6.3 to 11.5.56
        Workspaceb    423.5−7.0 to 14.1.5
        Integrationb    423.6−7.9 to 15.1.53
        Patient identificationb    423.1−3.6 to 9.8.36
        Patient engagementb    423.2−5.3 to 11.6.45

    • CARE = Consultation and Relational Empathy; DASI = Duke Activity Status Index; ED = emergency department; GAD = generalized anxiety disorder; GAD = 7-item Generalized Anxiety Disorder screener; GAIN = Global Appraisal of Individual Needs; MMAS = Morisky Medication Adherence Scale; PHQ = Patient Health Questionnaire; PHQ-9 = 9-item Patient Health Questionnaire; SUD = substance use disorder.

    • ↵a N varies depending on availability of covariates.

    • ↵b Greater score indicates better health.

  • Table 4.

    Adjusted Association Between Baseline Practice Integration (PIP Total Score) and Patient Outcomes at Baseline and Follow-Up (N = 42; All Practices)

    PROMIS-29 DomainBaselineFollow-Up
    Effect95% CIP ValueEffect95% CIP Value
    Anxiety−0.06−0.12 to 0    .05−0.05−0.09 to −0.01    .01
    Depression−0.04−0.10 to 0.01    .1−0.04−0.08 to 0.01    .13
    Fatigue−0.03−0.07 to 0.01    .1−0.03−0.07 to 0.01    .19
    Sleep disturbance−0.05−0.10 to −0.01    .03−0.02−0.07 to 0.02    .3
    Pain interference−0.01−0.06 to 0.04    .64−0.03−0.07 to 0.02    .3
    Pain intensity−0.01−0.03 to 0.01    .52−0.01−0.02 to 0.01    .45
    Social participationa  0.05  0.01 to 0.09    .02  0.06  0.02 to 0.10    .01
    Physical functiona  0.04−0.01 to 0.08    .12  0.02−0.03 to 0.07    .39
    Physical health summarya  0.04−0.01 to 0.09    .12  0.02−0.03 to 0.07    .35
    Mental health summarya  0.05  0 to 0.09    .05  0.04  0 to 0.09    .04

    • PIP = Practice Integration Profile; PROMIS-29 = Patient-Reported Outcomes Measurement Information System.

    • ↵a Greater score indicates improved function.

Additional Files

Back to top

In this issue

Print
Download PDF
Article Alerts
Email Article
Citation Tools
Get Permissions

Jump to section

Subjects

Keywords