People with a mental illness have higher rates of chronic disease morbidity and mortality and a reduced life expectancy compared with the general population (
1–
3). Although the causes underlying such disparities are varied (
4–
7), a significant contributor is greater prevalence of modifiable health risk behaviors, including smoking (
8,
9), harmful alcohol consumption (
10), inadequate nutrition (
11,
12), and physical inactivity (
13,
14). Mental health care guidelines recommend the provision of preventive care to modify such client risk behaviors (
9,
15,
16). Recommendations include the “2As and R” model at a minimum, which entails assessment (A) of clients’ health risk behaviors and, for clients with risks, the provision of advice (A) and referral (R) to behavior change services (
17–
20).
Limited research has examined the provision of preventive care by mental health services for chronic-disease health risk behaviors, with such research predominantly focusing on smoking cessation care (
21–
26). Four studies within the past decade have examined the provision of care for more than one of these health risk behaviors (
27–
30). However, two have been limited to examining a single element of preventive care (
27,
29). One study addressed the provision of three elements of preventive care across four behaviors (smoking, harmful alcohol consumption, inadequate nutrition, and physical inactivity) (
30). The study, which surveyed clinical staff within Australian community mental health services, found that the proportion of clinicians who reported providing care to most (≥80%) of their clients for each of the four behaviors ranged from 13% to 89% for assessment, 46% to 80% for advice, and 23% to 61% for any type of referral. Care provision was consistently lowest for nutrition, and, for all behaviors, referral or follow-up was least likely to be provided (
30). An important limitation of studies that have examined the provision of preventive care to address a range of health risk behaviors in mental health services over the past decade is their use of either staff report (
28–
30) or medical record audit (
27). No studies, to our knowledge, have used client-reported receipt of preventive care despite its being suggested as a more appropriate measure (
31).
Previous research undertaken in the mental health setting has identified a number of clinical and client characteristics that are associated with greater provision of preventive care, including rural as opposed to urban location (
32,
33); a consultation of longer rather than shorter duration (
32); being seen by a nurse or allied health clinician as opposed to a psychiatrist (
32); diagnoses of diabetes, hypertension, obesity (
32,
33), asthma, or other respiratory disorders (
24); and psychiatric diagnoses, including bipolar disorder (
33) and some substance use disorders (
24).
A growing body of evidence suggests that mental health service clients are interested in improving their health risk behaviors (
34–
40) and suggests that mental health clients are receptive to smoking cessation care (
41) and care to increase physical activity (
34). However, client acceptance of preventive care for alcohol consumption or inadequate nutrition or of specific elements of preventive care has not been investigated. In the context of mental health clinicians perceiving that lack of client acceptance constitutes a barrier to provision of preventive care (
42–
44), there is a need for more research across a broader set of risk behaviors and elements of preventive care.
A study was undertaken within a network of Australian community mental health services to examine client-reported acceptance of receiving assessment, brief advice, and referral or follow-up from community mental health clinicians for each of four health risk behaviors; client-reported receipt of such forms of care for each of the four behaviors; and associations between client diagnosis, number of clinical appointments, and reported receipt of preventive care.
Methods
Design
A cross-sectional survey of community mental health service clients was undertaken from December 2011 through November 2012 in one local health district in New South Wales (NSW), Australia. Since 2010, the district has had a policy requiring community mental health clinicians to provide, in accordance with the 2As and R model (with assessment, advice, and referral constituting preventive care), routine preventive care to all clients for the four behavioral risks included in this study. At the time of the study, no specific training regarding the policy had been provided to clinicians (
45).
Ethical approval for the study was obtained from the Hunter New England Human Research Ethics Committee (approval no. 09/06/17/4.03) and the University of Newcastle Human Research Ethics Committee (approval no. H-2010–1116).
Participants and Recruitment
Community mental health services.
In Australia, public community mental health services provide ambulatory care to approximately 350,000 clients each year (
46). Within the study area, all 12 community-based mental health services providing care to adult clients were invited to participate. Such services receive over 7,000 new clients per year and provide general adult mental health care and care to more specialized populations, including services specializing in care for older persons, persons needing psychiatric rehabilitation, and those with an early psychiatric diagnosis, comorbid substance abuse, an eating disorder, or borderline personality disorder.
Community mental health clients.
Clients attending any of the 12 eligible services were initially eligible if, on the basis of electronic medical record data, they were at least 18 years of age, had attended at least one face-to-face appointment at an eligible service during the previous two weeks, had not been selected to participate previously, and had not been identified by their clinician as inappropriate to contact.
Over 12 months, a random sample of approximately 22 such clients (approximately 5% of weekly eligible clients) was selected weekly from the electronic medical records system using the survey select procedure in SAS, version 9.3. Selected clients were mailed an information letter and telephoned by trained interviewers to further determine eligibility, including speaking English, not living in a geriatric care facility or incarcerated, and being physically and mentally capable of responding to the survey items. Eligible consenting clients completed a computer-assisted telephone interview survey at that time or at a more suitable time.
Measures
Client descriptors.
Clients reported whether they were of Aboriginal or Torres Strait Islander origin (or both) and their highest education level, employment status, marital status, and psychiatric or general medical conditions for which they had received medical attention or taken medication within the previous two months. Age, gender, residential postal code, mental health service attended, and number of community mental health appointments attended within the past 12 months were obtained from electronic medical records for study consenters and nonconsenters.
Client health risk behaviors.
Clients reported the following during the month prior to seeing their community mental health clinician: tobacco smoking (
47), quantity and frequency of alcohol use (
48), fruit and vegetable consumption (
49), and physical activity (
50). Survey items are based on validated items from recommended assessment tools (
51–
54) and have been used in community surveys (
55). Clients’ risk status was based on Australian national guidelines (
47–
50).
Acceptability of preventive care.
Clients were asked to indicate the acceptability of clinician assessment for each behavior (strongly disagree to strongly agree). For example, “It is acceptable for [service] to ask you if you smoke any tobacco products.” Clients who were identified to be at risk of one or more behaviors were asked to similarly indicate the acceptability of clinicians’ providing brief advice and referral for addressing such behaviors.
Receipt of Preventive Care
Client receipt of assessment, brief advice, and referral or follow-up was assessed (possible responses were yes, no, and don’t know) (
17–
20). Clients were asked to report whether, during their community mental health appointments, the clinician assessed their smoking status, alcohol consumption, fruit and vegetable intake, and physical activity. Clients classified as at risk for a behavior and who reported being assessed for that behavior were asked whether their clinician advised them to modify their behavior. Similarly, clients classified as at risk for a behavior and who reported receiving an assessment for that behavior were asked whether their clinician spoke to them about or offered to arrange referral to the NSW Quitline for smoking (
www.icanquit.com.au/further-resources/quitline) or the NSW Get Healthy Coaching and Information Service for inadequate intake of fruits or vegetables or for physical inactivity (
www.gethealthynsw.com.au). No equivalent free, government-funded telephone helpline was available for reducing alcohol consumption. Such clients were further asked whether they were advised to see their general practitioner or Aboriginal Medical Service (AMS), or other services (such as a dietician, support groups, or Alcoholics Anonymous).
Analyses
Analyses were undertaken with SAS, version 9.3. Descriptive statistics were used to describe the sample characteristics and client acceptability of each element of care for each health risk behavior. Residential postal code was used to calculate each client’s geographic remoteness (
56) and socioeconomic index of disadvantage (
57). Condensed response categories were created for age, marital status, highest education level, geographic remoteness, index of disadvantage, and number of community mental health appointments in the previous 12 months. Chi square analysis was used to compare client groups—those consenting versus not consenting to the study—on age, gender, remoteness, and number of appointments. For chi square and regression analyses, variables were dichotomized (less than high school versus completed high school or greater).
For each behavior, a variable was created to reflect whether at-risk clients received any type of referral or follow-up. For clients who reported being at risk and being assessed, a “complete care” variable was created for receipt of both advice and referral or follow-up for each behavior (yes-no response).
Descriptive statistics were used to describe client receipt of assessment (responses were yes, no, or don’t know, with the latter two categories combined for analysis) for each behavior and to describe client receipt of advice, each type of referral, and complete care for each behavior for which a client was at risk and reported being assessed.
For each of the four behaviors, chi square analysis was used to examine the association between client diagnostic characteristics, the number of appointments in the past 12 months, and receipt of assessment and complete care. Variables associated at p≤.25 were entered into backward stepwise logistic regression models to examine their association with clients’ having received assessment and having received complete care for each of the four behaviors. [A table available in the
online supplement to this article provides detail on the chi square analyses.] The choice of this p value follows recommendations that more traditional levels of p<.05 often fail to identify important and clinically relevant variables (
58). Logistic regression models adjusted for age, gender, employment status, marital status, education, Aboriginal background, and residential remoteness in order to examine whether care receipt was independently associated with diagnostic characteristics and with the number of appointments in the past 12 months.
Results
Participants
All 12 community mental health services in the district participated. Of 1,106 clients selected to participate, 903 (82%) were contactable, and of these 129 (14%) were identified as ineligible for participation. Of the remaining 774 clients, 558 (72%) participated in the interview. A greater proportion of women compared with men agreed to participate (76% versus 68%, p=.009). There were no other significant differences between consenters and nonconsenters. Client descriptors are reported in
Table 1.
Acceptability of Receiving Preventive Care
Clients consistently reported that receiving preventive care would be highly acceptable (
Table 2). Acceptability of receiving an assessment ranged from 90% for assessment of fruit or vegetable consumption to 97% for assessment of alcohol consumption; acceptability of receiving advice ranged from 86% for smoking to 94% for physical inactivity; and acceptability of receiving a referral ranged from 88% for inadequate fruit or vegetable consumption to 91% for harmful alcohol consumption.
Receipt of Preventive Care
Assessment.
A majority of participants reported being assessed for smoking (73%) and alcohol consumption (76%). Over half were assessed for physical activity (57%), and a quarter for fruit or vegetable intake (26%) (
Table 3).
Advice.
Of participants who were at risk for a behavior and who reported receiving an assessment, a majority received brief advice to quit smoking (79%), reduce their alcohol consumption (73%), increase their consumption of fruits and vegetables (69%), and increase their physical activity (85%) (
Table 3).
Referral and follow-up.
Receipt of each type of referral or follow-up was low. Receiving information about telephone helplines ranged from 12% (physical inactivity) to 41% (smoking), whereas receiving an offer to arrange a referral to a helpline ranged from 1% (concerning fruit and vegetable consumption) to 7% (concerning smoking). Receipt of advice to speak to their general practitioner or AMS ranged from 2% (fruit and vegetable consumption and physical inactivity) to 13% (smoking), and being advised to use any other type of support ranged from 29% (smoking) to 41% (physical inactivity) (
Table 3).
Complete care.
For each of the four risk behaviors, less than half of the participants who were at risk and who reported being assessed received both advice and referral or follow-up, with the range from 37% (harmful alcohol consumption or inadequate fruit or vegetable consumption) to 48% (smoking).
Associations With Receipt of Preventive Care
Participants were less likely to be assessed for alcohol consumption if they had a diagnosis of schizophrenia (odds ratio [OR]=.63). Clients were more likely to be assessed for physical activity if they had more than two appointments in the past 12 months (3–11 appointments, OR=2.43; ≥12 appointments, OR=3.05). No significant associations were identified for assessment of smoking or consumption of fruits and vegetables (
Table 4).
If clients smoked, they were more likely to receive complete care when they had three to 11 appointments (OR=2.28) or 12 or more appointments (OR=2.88) in the past 12 months. Complete care for alcohol was more likely when a client had 12 or more appointments in the past 12 months (OR=3.19) and less likely if the client had a diagnosis of schizophrenia (OR=.21). Complete care for inadequate fruit or vegetable consumption was more likely when participants had diabetes (OR=4.22) or a respiratory condition (OR=3.32). Complete care for physical inactivity was more likely when a participant had diabetes (OR=4.59) (
Table 4).
Discussion
This study was the first to examine acceptability and receipt of preventive care for multiple health risk behaviors, as reported by mental health clients. Preventive care was highly acceptable to clients across all behaviors and care elements. Receipt of preventive care was variable across behaviors and care elements and was particularly low for inadequate nutrition, receipt of referral and follow-up, and receipt of complete care. Factors identified as being positively associated with preventive care receipt were having a greater number of appointments in the previous 12 months, a diagnosis of diabetes or a respiratory condition, and not having a diagnosis of schizophrenia.
The finding that the receipt of assessment, advice, and referral for all risk behaviors was highly acceptable to clients is consistent with research regarding smoking and physical activity (
34,
41) and extends previous findings to include acceptability for alcohol and nutrition and for referral. Such findings also extend the research identifying that a significant proportion of mental health service clients are interested in changing their health risk behaviors (
34–
40), suggesting that clinician beliefs regarding client nonreceptivity to general health care (
42,
43,
59) may be unfounded. The dissemination of training and educational resources has been reported to positively affect primary care nurses’ misconceptions regarding general medical care for clients with a mental illness (
60). The effectiveness of such strategies in reducing clinician misconceptions regarding client receptivity to preventive care and the impact on clinician provision of such care should be examined.
Although client reports indicated high acceptance of preventive care, receipt of care was low. The results are consistent with clinician reports of suboptimal preventive care provision, with care particularly low for inadequate consumption of fruits and vegetables and for referral across all behaviors (
30). For each of the four behaviors, less than half of participants who were at risk and reported receiving assessment were advised to use any type of referral or follow-up. Along with a perception that clients may not be receptive to or interested in preventive care, the low levels of referral may reflect poor communication between mental health and other health services (
61,
62), a perceived lack of referral options (
43,
61,
63,
64), or other organizational barriers. Further research is warranted to better understand the barriers to care provision in this setting in order to develop interventions to improve care.
For some behaviors and care elements, an association was identified between care receipt and the following: a greater number of appointments in the previous 12 months and a diagnosis of schizophrenia, diabetes, or respiratory illness. Such findings suggest that mental health clinicians are more likely to provide preventive care where they feel that time permits (
28,
44,
61) or when it is clinically indicated (
24,
33). Given that people with schizophrenia are at highest risk of chronic disease morbidity and mortality (
65,
66) and of experiencing a reduced life expectancy (
1,
3,
67), the finding that clients with a diagnosis of schizophrenia were less likely to receive care for alcohol consumption suggests that initiatives to increase the provision of such care for this client group are a particular priority. To maximize the benefits, it is important to provide preventive care routinely to all clients. Approaches to care delivery that limit the time required of clinicians, including reduced models of care, such as the 2As and R model, should be considered in the mental health setting (
17–
20). Further, systems changes such as information technology approaches to prompt preventive care delivery (
18,
68) and the automation of referrals should be implemented to support clinicians (
20,
69).
The results should be interpreted in light of a number of limitations. Client reports of receipt of preventive care in general health care settings have been acknowledged as more reliable than clinician reports (
31). However, to our knowledge no studies have reported the validity of such measures in mental health settings specifically. The extent to which the receipt of such care in this study is either an overestimate or an underestimate of the care actually received is unknown. Participants were selected on the basis of having a community mental health appointment in the prior two weeks. Because the survey questions addressed care without specification of time frame, some clients might have responded regarding the receipt of care over a longer period. Subsequent analyses have indicated that over 80% of clients responded to this item in terms of their most recent appointments with the service. Diagnoses were self-reported by participants and hence may reflect self-diagnosis rather than diagnosis by a health professional. Last, although data were collected from a health district covering a large geographical area with metropolitan, regional, and rural communities, the generalizability of the findings to other settings is unknown.
Conclusions
This study has demonstrated that although mental health service clients reported that preventive care for health risk behaviors is highly acceptable to them, their reported receipt of such care during community mental health appointments indicates that care is suboptimal. Poor physical health within this population is well documented, as is the high prevalence of health risk behaviors. Therefore, it is imperative that mental health services provide preventive care routinely. Strategies to increase the delivery of routine preventive care within mental health services, such as information technology approaches and automated referrals, are likely to be required.
Acknowledgments
The authors thank all members of the Preventive Care team, the electronic medical records team, the computer-assisted telephone interviewers, the Aboriginal Advisory group, and the community mental health services and clients for their contributions to the project.