Integrating behavioral healthcare for individuals with serious mental illness: A randomized controlled trial of a peer health navigator intervention

SCHRES-07020; No of Pages 7 Schizophrenia Research xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Schizophrenia Research journal homepage: www.elsevier.com/locate/schres

Integrating behavioral healthcare for individuals with serious mental illness: A randomized controlled trial of a peer health navigator intervention Erin Kelly a,b,⁎, Lei Duan a, Heather Cohen a, Holly Kiger a, Laura Pancake c, John Brekke a a b c

School of Social Work, University of Southern California, 669 West 34th Street, Montgomery Ross Fisher Building, Los Angeles, CA 90089, USA Health Services Research Center, University of California, Los Angeles, Los Angeles, CA 90023, USA Pacific Clinics, 2550 E. Foothill Blvd., Pasadena, CA 91107, USA

a r t i c l e

i n f o

Article history: Received 13 June 2016 Received in revised form 17 October 2016 Accepted 19 October 2016 Available online xxxx Keywords: Integrated health care Serious mental illness Intervention Peer

a b s t r a c t Objective: Individuals with serious mental illness also have high rates of comorbid physical health issues. To address those issues, this population needs interventions that improve self-management of health and healthcare. Methods: In order to improve the health and healthcare of individuals with serious mental illnesses, 151 consumers with serious mental illness were randomized to receive either usual mental healthcare plus the Bridge intervention (n = 76) or usual mental healthcare while on a 6 month waitlist (n = 75). The waitlist group received the intervention after the waitlist period. Results: Change score comparisons (difference of differences) of the treatment vs the waitlist groups revealed that the treated group showed significantly greater improvement in access and use of primary care health services, higher quality of the consumer-physician relationship, decreased preference for emergency, urgent care, or avoiding health services and increased preference for primary care clinics, improved detection of chronic health conditions, reductions in pain, and increased confidence in consumer self-management of healthcare. Conclusions: Peer providers using a manualized intervention can be an important part of the efforts to address the general medical care of individuals with serious mental illnesses. © 2016 Elsevier B.V. All rights reserved.

1. Introduction The physical health of individuals with serious mental illnesses is severely compromised. Across systematic reviews, there is evidence that individuals with a serious mental illness (SMI) are dying about 10– 20 years before their non-mentally ill peers (Chesney et al., 2014; De Hert et al., 2011; Walker et al., 2015). There are many reasons for this early mortality but largely it is due to preventable and treatable physical health conditions that are more prevalent and under-treated in this population. Individuals with SMI have higher rates of multiple chronic health conditions, such as diabetes, high blood pressure, high cholesterol, obesity, viral hepatitis, chronic obstructive pulmonary diseases, and cancer (Weber et al., 2009). These conditions are critical to address not only due to reduced quality of life and early mortality but because the cost of care for chronic conditions is also increasing rapidly. In 2002, treatment of chronic conditions cost approximately $331.9 billion but by 2013, costs had nearly doubled to $623.8 billion (Mandros, 2016). The higher rates of these conditions among this population are partly attributable to poor healthcare and lifestyle habits, but largely ⁎ Corresponding author. E-mail address: [email protected] (E. Kelly).

result from taking psychoactive medications and disparities in healthcare on system, provider, and patient levels (De Hert et al., 2011). It is critical to address the medical care factors that impair effective treatment of the physical health of individuals with SMI. There are numerous strategies that are being employed to coordinate the care of this population but only a few include self-management by consumers as a critical ingredient in their interventions (Kelly et al., 2014a). Across these many care integration models (Gerrity, 2016), an activated patient who can navigate a productive relationship with care providers is necessary. The “Bridge” is a comprehensive, healthcare engagement and selfmanagement intervention that teaches participants the skills to improve healthcare access and use. Our intervention is guided by Gelberg et al.’s (2000) Model for Vulnerable Populations, which includes the multitude of factors that can suppress or facilitate healthcare service use among those with SMI. The Bridge intervention has been described in detail previously (Brekke et al., 2013; Kelly et al., 2014b). Briefly, “the Bridge” intervention targets factors that negatively impact healthcare access, utilization, and outcomes among individuals with SMI. Consumers are taught the skills to access and manage their healthcare effectively by mental health peers known as peer health navigators. Peers are individuals who use their lived experience with

http://dx.doi.org/10.1016/j.schres.2016.10.031 0920-9964/© 2016 Elsevier B.V. All rights reserved.

Please cite this article as: Kelly, E., et al., Integrating behavioral healthcare for individuals with serious mental illness: A randomized controlled trial of a peer health navi..., Schizophr. Res. (2016), http://dx.doi.org/10.1016/j.schres.2016.10.031

2

E. Kelly et al. / Schizophrenia Research xxx (2016) xxx–xxx

recovery from mental illness, along with skills from formal training, to provide behavioral health services. Peers are a rapidly increasing part of the workforce (Bachrach et al., 2016) but only a few studies have included self-management skills training from peer-delivered interventions for physical health in this population; however, there are promising signs for the efficacy of interventions that include these components (Kelly et al., 2014a). Participants are empowered through training to increase their use of routine health services and screenings, which may lead to the detection of new health diagnoses, to develop improved relationships with health providers, and to increase their self-management of healthcare.

of time, billing, scope of practice, and documentation. According to staff and supervisors these challenges remained for the agency staff throughout the study. Second, we described the study design in detail to staff so that they knew that all participants would receive the intervention during the study period. Third, we asked supervisory staff and health navigators throughout the study if usual care had changed to mimic the intervention, which they did not report as occurring. Fourth, we instructed our health navigators not to share their intervention materials with the clinic staff, and they reported that they were not asked by clinic staff for these materials. 2.5. Measures

2. Methods 2.5.1. Character of the intervention 2.1. Setting The study was conducted in a large community mental health agency in California. The agency provides outpatient rehabilitative services to adults with SMI.

2.5.1.1. Peer providers. The three peer health navigators had caseloads of about 20 each throughout the study. Two of the navigators were African American females and one was a bilingual Latino male. Their lived experience came from personal recovery and/or the recovery of a loved one in their lives.

2.2. Sample The sample was recruited using a short screening form (6 items) designed to assess whether individuals were connected to medical care or had unaddressed medical issues. An affirmative response to any of the items was a positive screen. Referrals came from treatment teams using their existing mental health caseloads, or from information sessions held for agency consumers. Study inclusionary criteria were: 1) over the age of 18; 2) admitted to one of the programs at the study site; 3) local residence for at least 3 months; 4) English fluency, 5) capacity to give informed consent; 6) diagnosed with schizophrenia, schizo-affective disorder, bipolar disorder, or major depression. The exclusion criteria were: 1) under conservatorship; 2) unable to give informed consent; 3) currently hospitalized. Participants received insurance coverage through Medicaid (97%) or Calworks (3%). 2.3. Design Once consented, subjects were randomized using a computer-generated random number table to immediate health navigation or to a six-month waitlist, with health navigation commencing after the waitlist period for that group. 2.4. Procedures Data were collected in 3 waves with 6 month intervals between assessments, based on the Bridge intervention being designed to last six months. The same outcome measures were assessed at each time point in a face-to-face interview conducted by three trained research assistants. Several features of this design should be highlighted. First, the waitlist design ensured that all participants eventually received the full health navigation intervention. Second, the mental health services received were unchanged as a result of participating in the study. All participants received the array of mental health services to which they were entitled, including outpatient rehabilitative services that were field based or office based, and on-site psychiatric services. Third, we did not restrict the consumers who could participate in this study based on diagnosis, functional level, or medical history. Fourth, treatment teams in the waitlist condition were instructed to maintain their routine care with every consumer, including routine healthcare assistance. As such, the comparator was treatment as usual (TAU) Since this intervention was delivered within a single clinic, particular attention was paid to the issue of treatment contamination to TAU. First, we developed our intervention because the mental health staff (including psychiatric nurses) was unable to deal with health issues beyond those directly associated with mental health practice due to issues

2.5.1.2. Service engagement and working alliance. Participant engagement in the intervention was measured with the Service Engagement Scale based on navigator ratings (Tait et al., 2002). The Working Alliance Inventory short form (Hatcher and Gillaspy, 2006) measured the quality of the relationship from the participant's perspective. 2.5.1.3. Intervention fidelity and intensity. Intervention fidelity was measured using a 20-item instrument developed in our pilot work based on interview, role play, and case records. Navigators recorded the number, length, and nature of in-person contacts and phone calls where they spoke directly to the participants. 2.6. Health and healthcare measures 2.6.1. Health service utilization The preferred locus of care and health service use were assessed using two scales from an adapted version of the UCLA CHIPTS healthcare and health utilization survey (CHIPTS, 2012; Kelly et al., 2014b). First, participants identified where they usually seek care (emergency room, urgent care, primary care provider, clinic, or no place). Second, participants rated the frequency that they visited each type of provider (0 = never, 1 = once or twice, 2 = three to five times, 4 = over five times) in the prior 6 months. For analytic purposes, providers were classified as Emergency/Urgent Care if they were located in an emergency room or urgent care facility. Providers were classified as routine care providers if they were primary care, specialty care, dentists, optometrists, or alternative medicine practitioners. 2.6.2. Satisfaction with primary care provider Participants were asked if they had a primary care provider. Participants with a routine primary care provider completed the Engagement with the Healthcare Provider Scale (Bakken et al., 2000) regarding their relationship with their primary care provider. 2.6.3. Self-management attitudes and behaviors The ability to self-manage healthcare was evaluated for confidence and behaviors. Participants rated how confident they were about managing their health (1 = not at all confident to 10 = very confident) on a 10-item scale. Items were based on skills expected to develop in health navigation. The behavioral self-management scale was adapted from the Mental Health Confidence Scale (Carpinello et al., 2000). This 14item scale includes items on skills such as appointment making, pharmacy visits, establishing a medical home, and feeling that healthcare needs were heard and addressed. Participants estimated the frequency

Please cite this article as: Kelly, E., et al., Integrating behavioral healthcare for individuals with serious mental illness: A randomized controlled trial of a peer health navi..., Schizophr. Res. (2016), http://dx.doi.org/10.1016/j.schres.2016.10.031

E. Kelly et al. / Schizophrenia Research xxx (2016) xxx–xxx

3

Phase 1 In-person contact for 1-4 months of assessment, modeling, and coaching. Consumer becomes activated and gains confidence in ability to address health. Intervention Components: Screening and engagement Assessment and goal setting (healthcare and wellness goals) Preparing for the medical appointment Navigating the medical appointment (in vivo)

Phase 2 2-3 months of coaching, fading and reinforcing consumer selfmanagement. Consumer gradually assumes self-management of healthcare activities

Phase 3 Ongoing support and boosters as needed Consumer manages their health and healthcare to the highest degree that they can. Coordination with those who can assist in remaining or ongoing areas of need.

Intervention components: In vivo coaching and fading during healthcare activities Navigator support (in person or phone)

Reviewing the appointment and treatment/care plan (in vivo) Implementing care plan (in vivo)

Fig. 1. The phases of health navigation and intervention components.

of these experiences on a 3 point scale (1 = never to 3 = 3 or more times). 2.6.4. Routine health screening Participants were asked whether they had had an annual physical, dental, or eye exam in the prior 6 months (or since their last assessment). A count of routine screenings was created.

supervisor initially observed and participated in several sessions with consumers, provided modeling, coaching and detailed feedback and scored navigators' performance to optimize the fidelity of the intervention. Weekly supervision sessions were held with the navigators, the peer supervisor, and the PI throughout the study period. Table 1 Baseline descriptives of the full sample, treated and waitlist groups (n = 151).

2.7. Health status 2.7.1. Medical diagnoses A checklist of 10 chronic health diagnoses, (e.g., diabetes, high blood pressure, heart disease) was shown to participants and they identified if they had been diagnosed with each condition (ever at baseline, or in the last 6 months in subsequent assessments). 2.7.2. Pain Participants rated the severity of their pain and how much pain interfered with their functioning within the last week using 2 items drawn from the SF-12 (Ara and Brazier, 2008). A z-scored overall pain index was created from two items measuring pain and pain interference with functioning. 2.8. The intervention The “Bridge” intervention addresses care issues from prevention and screening to management of chronic conditions, although the specific medical services provided are determined by the off-site physician. Self-management is a critical feature of the intervention as many of the activities require active participation in one's healthcare. The intervention is manualized and uses motivational interviewing, cognitive behavioral strategies (modeling, coaching, reinforcement, role-playing, and fading) as well as psychoeducation about the healthcare system, benefits, health screenings, and working with medical providers. The intervention is delivered largely in vivo in community settings where physical health services are received, and is personalized to the healthcare experiences and needs of the participant. The goals are to increase access, improve the healthcare experience, and foster self-management. The phases and components of the six-month intervention are presented in Fig. 1. Training of the peer health navigators took place over two days using material from our health navigation certification course. The peer

Full sample M (SD) Gender Female Male Age

Immediate treatment

n

%

81 70

53.6% 46.4%

45.63 151 (10.95)

M (SD)

n

Waitlist %

35 46.1% 41 53.9% 44.80 76 (11.30)

M (SD)

n

%

46 61.3% 29 38.7% 46.47 75 (10.59)

Race/Ethnicity White Black Hispanic Other

37 12 90 12

24.5% 7.9% 59.6% 7.9%

17 7 46 6

22.4% 9.2% 60.5% 7.9%

20 5 44 6

26.7% 6.7% 58.6% 8.0%

Diagnosis Schizophrenia Schizoaffective Bipolar Depression Other

28 28 29 57 8

18.5% 18.5% 19.2% 38.5% 5.3%

18 14 14 25 5

23.7% 18.4% 18.4% 32.9% 6.6%

10 14 15 33 3

13.3% 18.7% 20.0% 44.0% 4.0%

91 1 39

60.3% 0.7% 25.8%

48 63.2% 1 1.3% 20 26.3%

43 57.3% 0 0.0% 19 25.3%

20

13.2%

7

13 17.3%

Income source SSI SSDI General relief/public assistance No income

Chronic medical disease diagnosis High blood 42 27.8% pressure High 40 26.5% cholesterol Lung disease 21 13.9% Diabetes 20 13.2% Hepatitis C 19 12.6%

9.2%

26 34.2%

16 21.3%

21 27.6%

19 25.3%

14 18.4% 9 11.8% 10 13.2%

7 9.3% 11 14.7% 9 12.0%

Note: Mental health diagnoses were drawn from electronic medical records. All other data in the table is self-reported by participants.

Please cite this article as: Kelly, E., et al., Integrating behavioral healthcare for individuals with serious mental illness: A randomized controlled trial of a peer health navi..., Schizophr. Res. (2016), http://dx.doi.org/10.1016/j.schres.2016.10.031

4

E. Kelly et al. / Schizophrenia Research xxx (2016) xxx–xxx

2.9. Statistical analyses

3.2. Sample retention

We examined intervention effects in the following steps: 1) change score comparisons from baseline to 6-month follow-up between the immediate treatment and waitlist control groups; 2) pre-intervention scores and post-intervention scores of the two groups were collapsed and within-person comparisons were completed to see whether outcomes improved over time; and 3) using the immediate treatment group alone, the 12-month follow-up scores were compared to 6month follow-up scores to see if the intervention effect sustained. We used independent t-tests, paired t-tests, or Chi-square tests to address the research questions. ANOVA comparisons tested for differences in response to the intervention across navigators and none were found. In addition, generalized linear mixed models (GLMM) were also performed and the results from both approaches were compared. We chose to present the results based on t-tests for ease of interpretation since both approaches have similar findings. Cohen's d was included as an estimate of effect size

Based on previous studies we projected a 30% attrition rate over the course of the study. The attrition rate was 19% at 6-months and 31% at 12 months (see Fig. 2). There were no demographic, navigator, or clinical characteristics that predicted participants' retention across the follow-up periods.

3. Results 3.1. Sample baseline characteristics Demographic and clinical characteristics of the full sample and immediate treatment and waitlist groups at baseline are presented in Table 1. There were no significant differences between groups at baseline for any demographic or clinical characteristics. In the 6 months prior to the study, 28% had not seen a routine care doctor and 42% had sought emergency/urgent care.

3.3. Intervention character and fidelity The Bridge intervention is personalized so the number of contacts varies according to need and engagement. The average number of inperson contacts during the first 6 months for the immediate intervention group was 4.91 (SD = 5.02) and the average number of phone calls where direct participant contact was made was 6.27 (SD = 6.70). The average meeting time for in person visits was 61 min (SD = 57 min) and for phone call contacts the average call lasted 3 min (SD = 13 min). The average service engagement score (range: 0–3) was 2.1 (SD = 0.86), and the average working alliance score (range 1–7) was 5.9 (SD = 1.3), suggesting that participants were well engaged, and developed positive working alliances with their peer navigators. Fidelity was rated after all three peer navigators had been active in the field for over 4 months. Using two raters (who were intervention trainers) the average fidelity scores were above the “good” range. 3.4. Between group comparisons Using an intent-to-treat approach, change score comparisons during the 6 months of the intervention were conducted between

Assessed for eligibility (n=189)

Excluded (n=38) Not meeting inclusion criteria (n=14) Declined to participate (n=24)

Enrollment

Randomized (n= 151)

Allocation Allocated to intervention (n= 76)

Allocated to waitlist (n=75)

Follow-Up Lost to follow-up 6 months (n= 15) Reasons: discontinued intervention (n =6), discontinued at agency (n=7), deceased (n=1), incarcerated (n =1)

Lost to follow-up 6 months (n= 13) Reasons: uninterested in intervention (n =3), discontinued at agency (n =8), deceased (n= 1), incarcerated (n =1)

Lost to follow-up 12 months (n=23) Reasons: discontinued intervention (n=12), discontinued services at agency (n=9), deceased (n=1), incarcerated (n =1)

Lost to follow-up 12 months (n=25) Reasons: uninterested in intervention (n = 11), discontinued at agency (n =11), deceased (n= 2), incarcerated (n =1)

Analysis Analysed Intent to Treat (n =61)

Analysed Intent to Treat (n = 62)

Available cases: 6 months (n = 61) 12 months (n = 53)

Available cases: 6 months (n = 62) 12 months (n = 50)

Fig. 2. CONSORT flow diagram.

Please cite this article as: Kelly, E., et al., Integrating behavioral healthcare for individuals with serious mental illness: A randomized controlled trial of a peer health navi..., Schizophr. Res. (2016), http://dx.doi.org/10.1016/j.schres.2016.10.031

E. Kelly et al. / Schizophrenia Research xxx (2016) xxx–xxx

the immediate treatment and waitlist groups and are presented in Table 2. Changes in health services were measured in terms of the preferred locus of care, where care was actually received, and relationship quality with their primary care provider. In order to test changes in preferred care locus at 6-month follow-up, participants were coded dichotomously as either retaining/developing preference to primary care clinics, or as not preferring primary care clinics or lost their preference for primary care clinics. In Chi-square comparisons, those in the immediate treatment group were significantly more likely to stay connected or become connected to primary care (80%) than those in the waitlist group (63%), after 6 months of the intervention. At baseline, there were no differences in the rates that participants in the treatment and waitlist groups saw routine care providers or how they rated the quality of their relationship with their primary care provider. However, participants in the immediate treatment group increased their visits to routine care providers more than the waitlist group at 6-month follow-up and they reported higher quality relationships with their primary care providers. Self-management of healthcare was examined in terms of attitudes (confidence) and behaviors. At baseline, individuals in the waitlist group were significantly more confident in their ability to self-manage

5

their healthcare; however, at six month follow-up, the change in confidence was significantly higher among those in the treated group than among those in the waitlist group. There were no between group differences in ratings of self-management behaviors at baseline or after 6 months. As under-detection of existing health conditions is a concern for this population (De Hert et al., 2011), we anticipated that the intervention would lead to higher rates of diagnoses for the treatment group compared to the waitlist group, which was confirmed. There was a trend for more routine health screenings in the treated group as well. There were also significant reductions in the severity of bodily pain reported by those in the treated group compared to the waitlisted TAU group.

3.5. Within group changes 3.5.1. Collapsed treatment effect Table 3. After 6 months of the intervention there were improvements in the detection of chronic diseases, reductions in pain, improved relationships with primary care providers, more visits to routine care providers, more preference for primary care, more routine health screenings, and greater healthcare self-management confidence.

Table 2 Independent t-test comparisons of the immediate treatment and waitlist control groups for health issues, healthcare services, and self-management from baseline to 6-month follow-up. Immediate treatment group (n = 76) Variables

M

n

M

SD

n

t

p

d

0.39 0.38 0.37

76 61 61

1.44 1.43 −0.02

0.41 0.38 0.33

75 62 62

−1.146 1.245 2.546

0.254 0.216 0.012

0.46

0.77 0.52 0.81

60 54 44

1.76 1.80 0.08

0.81 0.90 0.75

64 50 46

−0.443 −3.008 −2.396

0.659 0.004 0.019

−0.51

0.53 0.51 0.53

76 61 61

1.33 1.40 0.02

0.45 0.54 0.52

75 62 62

0.359 −0.450 −0.343

0.720 0.655 0.732

−0.06

0.90 1.04 1.10

76 61 61

0.76 0.95 0.15

0.89 0.95 0.94

75 62 62

0.649 2.092 1.788

0.517 0.038 0.076

0.32

PCP/Clinic 51 (67%) 49 (80%)

76 61

None/ER 23 (31%) 23 (37%)

PCP/Clinic 51 (69%) 39 (63%)

74 62

χ2 0.057 4.586

0.862 0.045

3.20 2.69 1.65

76 61 61

8.20 7.60 −0.28

2.44 3.03 1.78

75 62 62

−2.795 0.273 1.986

0.006 0.786 0.049

0.36

0.54 0.43 0.44

70 61 55

1.96 2.00 0.02

0.49 0.51 0.55

75 62 58

−0.99 −0.19 1.172

0.322 0.842 0.244

0.08

1.30 1.41 0.23

1.58 1.56 0.69

76 61 61

1.00 1.02 −0.08

1.23 1.32 0.91

75 62 62

1.312 1.508 2.123

0.191 0.134 0.036

0.38

3.55 3.25 −0.2

1.54 1.69 1.47

76 61 61

3.56 4.06 0.32

1.6 1.48 1.40

75 62 62

−0.029 −2.856 −2.007

0.977 0.005 0.047

−0.36

−0.02 −0.22 −0.16

0.95 0.92 0.75

76 61 61

0.02 0.22 0.11

0.92 0.89 0.86

75 62 62

−0.237 −2.652 −1.808

0.813 0.009 0.073

−0.33

Health services Number of routine care visits Baseline 1.36 Follow-up 1.51 Change score 0.14 Relationship quality with PCP Baseline 1.70 Follow-up 1.36 Change score −0.32 Number of ER/urgent care visits Baseline 1.36 Follow-up 1.35 Change score −0.01 Number of routine health screenings Baseline 0.86 Follow-up 1.33 Change score 0.48 Preferred locus of care Baseline Follow-up

None/ER 25 (33%) 12 (20%)

Self-management Confidence for self-management Baseline 6.91 Follow-up 7.74 Change score 0.33 Behavioral self-management Baseline 1.88 Follow-up 1.99 Change score 0.06 Health issues Chronic diseases Baseline Follow-up Change score Bodily pain severity Baseline Follow-up Change score Bodily pain index Baseline Follow-up Change score

Waitlist group (n = 75)

SD

Please cite this article as: Kelly, E., et al., Integrating behavioral healthcare for individuals with serious mental illness: A randomized controlled trial of a peer health navi..., Schizophr. Res. (2016), http://dx.doi.org/10.1016/j.schres.2016.10.031

6

E. Kelly et al. / Schizophrenia Research xxx (2016) xxx–xxx

4. Discussion

Table 3 Paired t-tests of collapsed treatment effects of the full sample.

Health services Number of routine care visits Relationship quality issues with PCP Number of ER/urgent care visits Number of routine screenings

Preferred locus of care is PCP Preferred locus of care not PCP Self-management Confidence for self-management Behavioral self-management Health issues Chronic diseases Bodily pain severity Bodily pain index

Treatment time 1

Treatment time 2

M

M

SD

n

SD

t

p

d

1.42 0.39 111 1.51

0.39 −2.245 0.027

0.21

1.66 0.76 81

1.41

0.62 2.947

0.004

0.34

1.31 0.44 111 1.25

0.36 1.394

0.166

0.13

0.92 0.92 111 1.20

0.99 −2.569 0.012

0.24

% 64%

n 70

% 79%

n 86

χ2

p

36%

39

21%

23

22.90

b0.001

7.52 2.09 111 7.96

1.89 −2.669 0.009

0.26

2.07 0.47 102 2.09

0.40 −0.532 0.596

0.05

1.14 1.40 111 1.29 1.40 −1.996 0.048 3.74 1.52 111 3.44 1.68 2.189 0.031 0.07 0.92 111 −0.10 0.95 2.274 0.025

−0.19 0.21 0.22

3.5.2. Enduring and delayed changes The immediate treatment group was assessed 6 months post-treatment in order to evaluate if the effects of the intervention were retained and to detect delayed effects. There were no statistically significant declines in any of the health or healthcare use indicators for the immediate treatment group from the end of the intervention to 6 months post-intervention (see Table 4). However, regarding delayed and continuing improvements, the immediate intervention group reported significantly fewer emergency room/urgent care visits and significantly more behavioral self-management compared to their 6month assessment, and over 91% (up from 80%) reported primary care as their preferred locus of care.

Table 4 Within-person mean change scores of the immediate treatment group for health issues, healthcare services, and self-management from 6-month to 12-month follow-up. Time 2 (6 months) to time 3 (12 months) (n = 61) Variables

M

SD

N

t

p

d

Health services Number of routine care visits Relationship quality issues with PCP Number of ER/urgent care visits Number of routine screenings

0.04 −0.05 −0.17 0.17

0.35 0.66 0.49 0.97

54 42 54 54

−0.864 −0.515 2.522 1.267

0.392 0.609 0.015 0.211

0.12 0.08 0.35 0.15

Preferred locus of care is PCP change

Total n 53

Δ% 9%

Δn 5

χ2 1.61

p 0.235

Self-management Confidence for self-management Behavioral self-management

0.29 0.13

1.67 0.35

54 48

1.265 2.494

0.211 0.016

0.17 0.37

Health issues Chronic diseases Bodily pain severity Bodily pain index

−0.02 0.17 0.19

1.00 1.65 0.91

54 53 53

0.136 −0.749 −1.876

0.892 0.457 0.066

0.02 0.10 0.26

Note: The change in the preferred locus of care reflects a change from 43 to 48 individuals who prefer the primary care provider clinic as their main source of care.

Improving the physical healthcare of individuals with serious mental illnesses is vital to their well-being and to decreasing their mortality. The Bridge intervention demonstrated improvements for the health and healthcare of individuals with serious mental illnesses in several important areas, and those gains remained six months after the intervention. There were also indications of delayed improvement in two key areas, self-management behaviors and reduced emergency room use. Self-management is a core principal of this intervention and key to maintenance of gains in health and healthcare use. The increased detection of previously undiagnosed chronic diseases in this study was likely due to persistent under-detection of diseases in this population (De Hert et al., 2011). The finding that self-management confidence increased while participants received more chronic disease diagnoses supports the viability of this intervention and the importance of having an active intervention for individuals with serious mental illnesses while receiving healthcare. Importantly, there was evidence of attitudinal and then behavioral changes in the self-management of health and healthcare by those in the treated group, which means that there was a logical progression of self-management adoption over time. It is also possible that changes in self-management behaviors need more emphasis earlier in the intervention. Improvements in the location, frequency, and quality of relationships with routine care were also noteworthy changes. The intervention helped consumers change where they would first seek medical care towards a primary care setting, increased their visits to routine providers, and improved their relationship quality with their provider, which are all important changes if we seek to improve healthcare and health outcomes. Decreasing use of emergency services took a longer time as there was a significant reduction in emergency service use by the intervention group in the six months post-intervention, which suggests that dependence on emergency services might need special attention for this population. Data on the intervention's character suggest that subjects were well integrated into the intervention and developed positive alliances with their peer navigators. Data on intensity suggest it is a modestly intensive intervention with about one in-person and one phone contact per month over 6 months, but with large individual variability. In supervision, navigators noted that some participants moved very quickly with navigation and others moved more slowly and needed more contact. The overall modest intensity of the intervention bodes well for its widespread utility. The agency also uses protocols for peers to bill Medicaid for these services. Given that the majority of participants in this study were ethnic minorities, particularly Latino, these findings have notable relevance to health disparities for ethnic minority populations who are also diagnosed with a mental illness. Systematic reviews on interventions targeting the health of those with serious mental illnesses have highlighted the need for interventions that are useful to racial/ethnic minorities (Siantz and Aranda, 2014; Cabassa et al., 2016), and this intervention shows promise for addressing this gap. In post hoc analyses not reported here we did not find evidence of a differential response to the intervention by race/ethnicity, but we found that Latinos received significantly more health screenings and accessed more routine care, but also reported less confidence in the self-management of healthcare than non-Latinos. We will report findings on ethnic minorities in a future manuscript. The data on the health and healthcare of participants in the present paper are from self-reports. In a subsequent paper we will evaluate whether insurance records support the improvements in health and healthcare and the impact of this intervention on healthcare costs. Health navigation is an individually tailored intervention and future research should explore whether particular subpopulations respond differentially to the intervention or whether specific strategies used by navigators were more strongly associated with improvements. We

Please cite this article as: Kelly, E., et al., Integrating behavioral healthcare for individuals with serious mental illness: A randomized controlled trial of a peer health navi..., Schizophr. Res. (2016), http://dx.doi.org/10.1016/j.schres.2016.10.031

E. Kelly et al. / Schizophrenia Research xxx (2016) xxx–xxx

also had no measures of health status such as laboratory and medical diagnostic tests, however, these outcomes are hard to establish in an intervention that seeks to improve overall health rather than a single chronic disease. Lab measures are important for future research but, recent analyses also attest to the importance of subjective measures of health status (Kelly et al., 2014a). A recent meta-analysis of peer delivered interventions found that they showed generally small to modest effect sizes (Fuhr et al., 2014). The current peer intervention fits this pattern. Given recent reports that advocate for the use of peers in mental health and health services, these findings can be seen as encouraging (Bachrach et al., 2016; Hardin et al., 2014; National Association of State Mental Health Program Directors, NASMHPD, 2014). Overall, the “Bridge” intervention is a promising peer-delivered intervention to address the healthcare and health needs of individuals diagnosed with serious mental illness. Funding acknowledgement This work was (partially) supported through a Patient-Centered Outcomes Research Institute (PCORI) Award (#6650). Disclaimer All statements in this report, including its findings and conclusions, are solely those of the authors and do not necessarily represent the views of the Patient-Centered Outcomes Research Institute (PCORI), its Board of Governors or Methodology Committee. Conflict of interest The University of Southern California and Pacific Clinics have a financial stake in the outcomes of this research. None of the researchers have a financial stake in these data. Acknowledgements We would like to thank our health navigators and their supervisor, Toni Jackson, Tamara Sankofa-Ra, Francisco Espinoza, and Lou Mallory. We would also like to thank Zoey Greer, Armando Ruan, Crystal Stewart, and Jorge Avila for all their efforts to make this project a success.

References Ara, R., Brazier, J., 2008. Predicting the short form-6D preference-based index using the eight mean short form-36 health dimension scores: estimating preference-based health-related utilities when patient level data are not available. Value Health 12 (2), 346–353. Bachrach, D., Guyer, J., Levin, A., 2016. Medicaid Coverage of Social Interventions: A Road Map for States. Milbank Memorial Fund.

7

Bakken, S., Holzemer, W.L., Brown, M.A., et al., 2000. Relationships between perception of engagement with health care provider and demographic characteristics, health status, and adherence to therapeutic regimen in persons with HIV/AIDS. AIDS Patient Care STDs 14 (4), 189–197. Brekke, J.S., Siantz, E., Pahwa, R., et al., 2013. Reducing health disparities for people with serious mental illness. Best Prac. Ment. Health 9 (1), 62–82. Cabassa, L.J., Camacho, D., Velez-Grau, C.M., Stefancic, A., 2016. Peer-based health interventions for people with serious mental illness: a systematic literature review. J. Psychiatr. Res. 84, 80–89. Carpinello, S.E., Knight, E.L., Markowitz, F.E., et al., 2000. The development of the mental health confidence scale: a measure of self-efficacy in individuals diagnosed with mental disorders. Psychiatr. Rehabil. J. 23 (3), 236–244. Center for HIV Identification, Prevention, and Treatment Services (CHIPTS), 2012r. Health Care Utilization Survey. Retrieved from. http://chipts.ucla.edu/wp-content/uploads/ downloads/2012/02/Healthcare-Utilization-Providers-STD-Preg.pdf. Chesney, E., Goodwin, G.M., Fazel, S., 2014. Risks of all-cause and suicide mortality in mental disorders: a meta-review. World Psychiatry 13 (2), 153–160. De Hert, M., Correll, C.U., Bobes, J., et al., 2011. Physical illness in patients with severe mental disorders. I. Prevalence, impact of medications and disparities in health care. World Psychiatry. 10 (1), 52–77. Fuhr, D.C., Salisbury, T.T., De Silva, M.J., et al., 2014. Effectiveness of peer-delivered interventions for severe mental illness and depression on clinical and psychosocial outcomes: a systematic review and meta-analysis. Soc. Psychiatry Psychiatr. Epidemiol. 49 (11), 1691–1702. Gelberg, L., Andersen, R.M., Leake, B.D., 2000. The behavioral model for vulnerable populations: application to medical care use and outcomes for homeless people. Health Serv. Res. 34 (6), 1273–1302. Gerrity, M., 2016. Milbank Report: Evolving Models of Behavioral Health Integration: Evidence Update 2010-2015. Accessed on May 12, 2016. http://www.milbank.org/ uploads/documents/Evolving%20Models%20of%20BHI.pdf. Hardin, P., Padron, J., Manderscheid, R., 2014. White Paper: US Peer Leadership & Workforce Development. Hatcher, R.L., Gillaspy, J.A., 2006. Development and validation of a revised short version of the Working Alliance Inventory. Psychother. Res. 16 (1), 12–25. Kelly, E.L., Fenwick, K.M., Barr, N., et al., 2014a. A systematic review of self-management health care models for individuals with serious mental illnesses. Psychiatr. Serv. 65 (11), 1300–1310. Kelly, E.L., Fulginiti, A., Pahwa, R., et al., 2014b. A pilot test of a peer navigator intervention for improving the health of individuals with serious mental illness. Community Ment. Health J. 50 (4), 435–446. Mandros, A., 2016. What Is The Cost Of Chronic Conditions To The U.S. Health Care System?: An OPEN MINDSMarket Intelligence Report. Retrieved on May 10, 2016 from. https://www.openminds.com/intelligence-report/what-is-the-cost-of-chronicconditions-an-open-minds-market-intelligence-report/. National Association of State Mental Health Program Directors (NASMHPD), 2014o. Enhancing the Peer Provider Workforce: Recruitment, Supervision and Retention. Siantz, E., Aranda, M.P., 2014. Chronic disease self-management interventions for adults with serious mental illness: a systematic review of the literature. Gen. Hosp. Psychiatry 36 (3), 233–244. Tait, L., Birchwood, M., Trower, P., 2002. A new scale (SES) to measure engagement with community mental health services. J. Ment. Health 11 (2), 191–198. Walker, E.R., McGee, R.E., Druss, B.G., 2015. Mortality in mental disorders and global disease burden implications: a systematic review and meta-analysis. JAMA Psychiat. 72 (4), 334–341. Weber, N.S., Cowan, D.N., Millikan, A.M., et al., 2009. Psychiatric and general medical conditions comorbid with schizophrenia in the National Hospital Discharge Survey. Psychiatr. Serv. 60 (8), 1059–1067.

Please cite this article as: Kelly, E., et al., Integrating behavioral healthcare for individuals with serious mental illness: A randomized controlled trial of a peer health navi..., Schizophr. Res. (2016), http://dx.doi.org/10.1016/j.schres.2016.10.031