J Clin Psychol Med Settings (2014) 21:10–18 DOI 10.1007/s10880-013-9376-x
Health-Related Outcomes Associated with Patterns of Risk Factors in Primary Care Patients Jennifer S. Funderburk • Stephen A. Maisto Allison K. Labbe
•
Published online: 25 October 2013 Springer Science+Business Media New York (outside the USA) 2013
Abstract It is important to find ways to identify prevalent co-occurring health risk factors to help facilitate treatment programming. One method is to use electronic medical record (EMR) data. Funderburk et al. (J Behav Med 31:525–535, 2008) used such data and latent class analysis to identify three classes of individuals based on standard health screens administered in Veterans Affairs primary care clinics. The present study extended these results by examining the health-related outcomes for each of these identified classes. Follow-up data were collected from a subgroup of the original sample (N = 4,132). Analyses showed that class assignment predicted number of diagnoses associated with the diseases that the health screens target and number of primary care behavioral health, and emergency room encounters. The findings illustrate one way an EMR can be used to identify clusters of individuals presenting with multiple health risk factors and where the healthcare system comes in contact with them.
J. S. Funderburk (&) S. A. Maisto A. K. Labbe VA Center for Integrated Healthcare, Syracuse Veterans Affairs Medical Center, 800 Irving Ave., (116C), Syracuse, NY 13210, USA e-mail:
[email protected] S. A. Maisto e-mail:
[email protected] A. K. Labbe e-mail:
[email protected] J. S. Funderburk S. A. Maisto A. K. Labbe Department of Psychology, Syracuse University, 430 Huntington Hall, Syracuse, NY 13244, USA J. S. Funderburk Department of Psychiatry, University of Rochester, Rochester, NY, USA
123
Keywords Multiple risk factors Primary care Veterans Latent class analysis Regression
Introduction Research has shown that factors such as physical inactivity, poor diet, tobacco use, heavy alcohol use, depression, and anxiety are related to the development and/or progression of cardiovascular diseases, hypertension, and diabetes (Critchley & Capewell, 2003; Haapanen, Miilunpalo, Vuori, Oja, & Pasanen, 1997; Jonas, Franks, & Ingram, 1997; Katon et al., 2008; Klatsky, 1996; Lloyd-Jones et al., 2009; Richardson, Egede, Mueller, Echols, & Gebregziabher, 2008; Strik, Denollet, Lousberg, & Honig, 2003). Tobacco use, risky alcohol use, physical inactivity, and poor dietary practices have been identified as contributing to mortality (McGinnis & Foege, 1993), and U.S. population-based survey studies have shown that a substantial proportion of the population (52.0–83.7 %) has at least two of these risk factors (Berrigan, Dodd, Troiano, Krebs-Smith, & Barbash, 2003; Coups, Gaba, & Orleans, 2004; Fine, Philogene, Gramling, Coups, & Sinha, 2004; Pronk et al., 2004). There are health behavior interventions that address each of these risk factors singly, rather than in combination. Examples include brief advice for risky alcohol use (Ockene, Adams, Hurley, Wheeler, & Hebert, 1999; World Health Organization Brief Intervention Study Group, 1996), counseling for smoking cessation (An et al., 2006), behavioral interventions to increase physical activity (Calfas et al., 1996), and high-intensity counseling for obesity (Ely et al., 2008). However, there is growing evidence that interventions targeting multiple risk factors can have greater impact on health outcomes and healthcare utilization than single behavioral interventions (Goldstein,
J Clin Psychol Med Settings (2014) 21:10–18
Whitlock, & DePue, 2004). For example, Hyman, Pavlik, Taylor, Goodrick, and Moye (2007) found that a simultaneous intervention was more effective for physical activity and diet than interventions delivered sequentially. It is believed that interventions targeting multiple behaviors simultaneously can more efficiently deliver the intervention and allow for discussion of the interrelationships among risk factors (Prochaska, 2008). Given that multiple health risk factors exist among a majority of Americans, it is essential to find ways to identify groups of patients reporting multiple risk factors to help examine current treatments, and whether there is a need for more tailored interventions for specific groups of patients. One way to identify groups of patients reporting multiple risk factors is through use of electronic medical records (EMRs), which provide clinicians with easily accessible and comprehensive patient information. These records are an ideal mechanism to detect patients who report multiple health risk factors and can be used to trigger specific provider actions as a result. Using latent class analysis (LCA), an empirical method that examines the interrelationships among risk factors and characterizes the underlying set of mutually exclusive latent classes that account for the observed relationships, Funderburk, Maisto, Sugarman, and Wade (2008) analyzed data from EMRs to define clusters of patients based on associations among risk factors identified by several standard health screens used regularly (at least annually) in Veterans Affairs (VA) primary care clinics, including those for elevated blood pressure, smoking, depression, risky alcohol use, and posttraumatic stress disorder (PTSD). Splitting the original sample into two random samples, Funderburk et al. (2008) conducted exploratory (N = 3,315) and confirmatory (N = 6,728) LCAs, which identified a 3-class model as the best fitting model. Individuals in Class 1 tended to be healthier with a low probability of scoring positive on multiple health screens, but a moderate-to-high probability of screening positive for elevated blood pressure. Individuals in Class 2 had the greatest probability of screening positive on more than one health screen, with a moderate-to-high probability of reporting risky alcohol use, smoking, depression, and posttraumatic stress. Generally, patients in Class 2 were most likely to screen positive on the most health screens. Class 3 consisted of individuals demonstrating a higher probability of scoring positive on three specific health screens, notably smoking, elevated blood pressure, and risky alcohol use, and a low probability of scoring positive on any of the other screens (Funderburk et al., 2008). Compared to the other classes, Class 1 appears to have the least need for intervention and therefore will be referred to in this paper as the ‘‘Low Treatment Need’’ group. Class 2 appears to have the highest need for intervention(s) targeting several risk factors and therefore will be referred to as the ‘‘High
11
Treatment Need’’ group. Finally, Class 3 appears to fall in between Class 1 and Class 2 in regards to level of intervention need and therefore will be referred to as the ‘‘Moderate Treatment Need’’ group. When a Veteran screens positive on any one of these screens, the VA has implemented standard recommended clinical practice guidelines to help providers understand how to proceed; however, these guidelines tend to be specific to each screen, and the interventions offered (e.g., the smoking cessation program) typically focus on a single risk factor. Gathering a better understanding for those Veterans within each of these classes may help to identify how often and where the healthcare system comes in contact with these specific groups of Veterans and the extent of the negative health outcomes associated with assignment to one of those groups. Therefore, the overall goal of this study was to build upon Funderburk et al.’s (2008) findings by investigating the health-related outcomes and healthcare utilization associated with each of the identified classes. The study’s first aim was to examine the association between each group of patients identified at Time 1 and their healthcare utilization during a follow-up period (Time 2). Understanding the potential contact healthcare providers have with each group of patients is extremely important. If a certain group of patients identified at Time 1 presents for treatment less frequently at Time 2, then research may be conducted to examine the barriers to treatment or to identify the healthcare setting most likely to encounter these patients. For example, if a certain group of patients identified at Time 1 presents for treatment more frequently within the emergency room, then it might suggest the need to investigate ways to help these individuals in a different setting to help reduce costs associated with emergency care. We hypothesized that individuals in the High Treatment Need and Moderate Treatment Need groups would have significantly more primary care encounters than individuals in the Low Treatment Need group. Also, because individuals assigned to the High Treatment Need group tended to have a higher number of positive screens, indicating a higher probability of experiencing symptoms related to several chronic diseases in need of care, we expected that individuals in the High Treatment Need group would have the highest number of primary care encounters. We also hypothesized that individuals in the High Treatment Need and Moderate Treatment Need groups would have a greater likelihood of encounters with behavioral health clinics and the emergency room than individuals in the Low Treatment Need group, and that individuals in the High Treatment Need group would have the greatest likelihood of having an encounter in both settings. The second aim of this study was to determine if the identified classes predicted the number of diagnoses related to the modifiable health factors for which patients were
123
12
J Clin Psychol Med Settings (2014) 21:10–18
screened. We predicted that individuals in the High Treatment Need and Moderate Treatment Need groups would have a significantly greater number of diagnoses than individuals in the Low Treatment Need group during the followup period, with people in the High Treatment Need group having the highest average number of diagnoses.
the presence of PTSD symptoms: re-experiencing, avoidance, hyperarousal, and numbing. Each item is dichotomous (Yes or No) with a total range of 0–4 points for the measure. A score of 3 or more for both men and women indicated potential PTSD (Bliese et al., 2008; Prins et al., 2003) and has been shown to have a specificity of 87 % and a sensitivity of 78 % (Prins et al., 2003).
Methods
Blood Pressure
The original data source utilized by Funderburk et al. (2008) was the Department of Veterans Affairs (DVA) EMR database for the Veterans Affairs Healthcare Network of Upstate New York (VISN 2). This database was searched to identify patients seen in a VISN 2 primary care clinic between January 1, 2005 and June 30, 2005 and had scores entered into their EMR on six specific health screen measures (N = 10,043 patients).
Following established clinical guidelines, blood pressure readings are taken at every primary care visit, and any systolic blood pressure reading above 119 and/or diastolic blood pressure above 79 may indicate risk of high blood pressure (U.S. Department of Health and Human Services, 2003).
Measures
Following established clinical guidelines, patients’ height and weight are measured at every primary care visit and any body mass index above 25 is identified as indicative of being overweight or obese (Institute for Clinical Systems Improvement, 2006).
Alcohol Use The 3-item Alcohol Use Disorders Identification TestConsumption (AUDIT-C; Bush, Kivlahan, McDonell, Fihn, & Bradley, 1998) is completed annually by primary care providers to assess the quantity and frequency of drinking alcohol. Each item is rated on a 0–4 point scale with a total range of 0–12 points for the measure. A score of 4 or more for men and 3 or more for women indicated potential at-risk drinking and has been shown to have sensitivity of 86 % and specificity of 72 % (Bush et al., 1998).
Body Mass Index
Tobacco Use Patients are asked annually whether they are current smokers, quit less than 1 year ago, quit more than 1 year ago, or never smoked. In accord with clinical guidelines, any patients reporting current smoking are considered atrisk (National Institute of Drug Abuse, 2006). Procedures
Depression Prior to the national implementation of the Patient Health Questionnaire-2 (Kroenke, Spitzer, & Williams, 2003) within the VA, the 12-item General Health Questionnaire (GHQ; Goldberg & Williams, 1978) was completed annually within VISN 2 to identify individuals experiencing symptoms of depression (Smit, Beekman, Cuijpers, de Graaf, & Vollebergh, 2004). Each item is rated on a 4-point scale (from 0 to 3) with a total range of 0–36 points for the measure. Items were summed and a cut-off of 2 was used to identify those in need of further assessment as that cut-off has been shown to have a sensitivity of 82 % and specificity of 77 % in a U.S. sample (Goldberg et al., 1997). Post-Traumatic Stress Symptoms The 4-item Primary Care-Post Traumatic Stress Disorder Screen (PC-PTSD; Prins et al., 2003) is used to screen for
123
For this study, follow-up outcome data were collected by taking a random sample (N = 4,965) from the original data source (see Figs. 1, 2 and 3 for each group’s LCA risk factor profile). The number of primary care, behavioral health, and emergency room visits was obtained from each patient’s EMR covering the time period between January 1, 2006 and March 1, 2007. Further, any new ICD-9 codes entered for each patient’s visit from January 1, 2006 until March 1, 2007 were also obtained. Five individual diagnosis variables were created, where data were coded as ‘‘1’’ if a patient had any new ICD-9 code given at any subsequent primary care encounter that corresponded to an alcohol use disorder (i.e., 303.0, 303.9, 305.0, 291.0– 291.9), depressive disorder (i.e., 300.4, 311, 296.2– 296.3), PTSD (309.81), tobacco use disorder (305.1), or cardiovascular/cerebrovascular disorder (i.e., 402–404.9, 410–416.9, 428–438.9, 441–443.9). Otherwise, data were coded as ‘‘0.’’ If a patient was given distinct yet
J Clin Psychol Med Settings (2014) 21:10–18 1
Conditional Probability
Fig. 1 The conditional probabilities of each positive risk factor in the low treatment need group
13
0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Alcohol Use
Depression
Smoking
Blood Pressure
Posttraumatic Stress
Positive Health Screen
1
Conditional Probability
Fig. 2 The conditional probabilities of each positive risk factor in the moderate treatment need group
0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Alcohol Use
Depression
Smoking
Blood Pressure
Posttraumatic Stress
Positive Health Screen
1
Conditional Probability
Fig. 3 The conditional probabilities of each positive risk factor in the High Treatment Need group
0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Alcohol Use
Depression
Smoking
Blood Pressure
Posttraumatic Stress
Positive Health Screen
overlapping diagnoses (e.g., MDD and Depression NOS) at different primary care encounters, data were entered such that the patient was given a ‘‘1’’ for depression diagnosis. Once those dichotomous variables were created, a summed variable for each patient was created to index the number of diagnoses related to these modifiable health factors targeted by the screening measures diagnosed during the follow-up period. All study procedures were approved by the Syracuse Veterans Affairs Medical Center Institutional Review Board. Linear regression analyses were used to test whether group predicted number of primary care encounters and number of diagnoses related to modifiable health factors. Due to highly skewed data as a result of many patients not having any behavioral health or emergency room
encounters during the time period, logistic regression analyses were used to test likelihood of having a behavioral health encounter and an emergency room encounter. Demographic variables (age, gender, marital status) served as covariates in all analyses. Assumptions for linear and logistic regression analyses were met.
Results Participants Data on the number of visits to primary care, behavioral health, and emergency room from January 1, 2006 to
123
14
J Clin Psychol Med Settings (2014) 21:10–18
March 1, 2007 were collected on 4,132 Veterans. Information on any new diagnoses could only be pulled if the Veteran was seen by a VA healthcare provider during that time period (n = 2,570); otherwise, the presence of any new diagnoses was unknown. As shown in Table 1, a majority of the sample was male (97 %), Caucasian (60 %), and married (65 %). Nearly two-thirds of the sample (62.4 %) was identified as belonging to the Low
Treatment Need group, 30 % were identified as belonging to the Moderate Treatment Need group, and 8 % were identified as belonging to the High Treatment Need group. These results are similar to those found in Funderburk et al. (2008). The most common new diagnosis was one that corresponded with a cardiovascular/cerebrovascular disorder (see Table 1). Regression Analyses
Table 1 Sample characteristics Total sample (N = 4,132) N (%)
Diagnosis sub-sample (n = 2,570) N (%)
Low treatment need
2,580 (62.4)
1,568 (61.0)
Moderate treatment need
1,223 (29.6)
765 (29.7)
329 (8.0)
237 (9.2)
3,998 (96.8)
2,466 (96.0)
Married
2,687 (65.1 %)
1,636 (63.7)
Othera
1,445 (35.0 %)
934 (36.3)
2,470 (60.0 %)
1,587 (61.8)
Class assignment
High treatment need Gender Male Marital status
Race/ethnicity Caucasian African-American Unknown/preferred not to answer Otherb
149 (3.6 %)
104 (4.1)
1,497 (36.2 %)
866 (33.7)
16 (0.4 %)
13 (0.51)
ICD-9 diagnoses Alcohol use disorder
67 (2.6)
Depressive disorder
98 (3.8)
PTSD
46 (1.79)
Tobacco use disorder
98 (3.8)
Cardiovascular/cerebrovascular disorder a
234 (9.1)
Other includes never married, divorced, widowed, and separated
b
Other includes Asian, native Hawaiian or other Pacific islander, American Indian or Alaskan Native
Table 2 Primary care encounters and severity scores
A linear regression analysis found that class assignment significantly predicted number of diagnoses related to the modifiable health factors (R2 = 0.019, F(2,2563) = 3.66, p \ 0.001). Group assignment accounted for a significant increase in variance in the outcome variable beyond what was contributed by the demographic variables (DR2 = 0.011, F(1,2565) = 28.21, p \ 0.001). Contrast analyses revealed that patients belonging to the High Treatment Need group had, on average, a significantly greater number of diagnoses than patients belonging to the other two groups (p’s \ 0.0001). The number of diagnoses did not significantly differ between patients belonging to the Low Treatment Need and Moderate Treatment Need groups (p [ 0.05; see Table 2). A linear regression analysis indicated that class assignment significantly predicted number of primary care encounters (R2 = 0.024, F(2,4132) = 12.9, p \ 0.0001) and accounted for a significant increase in variance in the outcome variable beyond what was contributed by the demographic variables (DR2 = 0.003, F(1,4127) = 12.5, p \ 0.001). Contrast analyses found that patients identified as belonging to the High Treatment Need group had, on average, significantly more primary care encounters than patients identified as belonging to the other two groups (all p’s \ 0.001). There was no significant difference between patients belonging to the Low Treatment Need and Moderate Treatment Need groups (p [ 0.05; see Table 2). Logistic regression analyses revealed that patients in the High Treatment Need group were 4.91 and 4.58 times more likely than patients in the Low Treatment Need group and M (SD)
F value (comparisons to Low treatment need group)
F-value (comparisons to moderate treatment need group)
Primary care encounters
Primary care encounters occurred between July 1, 2005 and March 1, 2007 ** p value \0.0001, * p value \0.001
123
Low treatment need
2.61 (2.03)
–
–
Moderate treatment need
2.66 (2.16)
0.61
–
High treatment need
3.24 (2.28)
15.27**
10.68*
Number of diagnoses Low treatment need
0.19 (0.44)
–
–
Moderate treatment need High treatment need
0.20 (0.46) 0.38 (0.64)
.68 27.59**
– 19.84**
J Clin Psychol Med Settings (2014) 21:10–18 Table 3 Behavioral health and emergency room encounters
15
n Had encounter
n No encounter
Odds ratio (OR), 95 % CI (comparison to low treatment need group)
Odds ratio (OR), 95 % CI (comparison to moderate treatment need group)
Behavioral health encounters Behavioral health and emergency room encounters occurred between January 1, 2006 and March 1, 2007. The n for behavioral health and emergency room encounters represents the number of individuals in each group that had at least one encounter * p value \0.05, ** p value \0.001, *** p value \0.0001
Low treatment need
145
2,435
Moderate treatment need
70
1,153
High treatment need
104
225
– 1.07 (0.79–1.45) 4.91 (3.62–6.68)***
– – 4.58 (3.21–6.53)***
Emergency room encounters Low treatment need
463
2,117
Moderate treatment need
202
1,019
High treatment need
164
165
Moderate Treatment Need group, respectively, to have a behavioral health encounter. Additionally, individuals in the High Treatment Need group were 2.96 and 3.25 times more likely than individuals in the Low Treatment Need group and the Moderate Treatment Need group, respectively, to have an emergency room encounter. People in the Moderate Treatment Need group were not more likely to have either behavioral health or emergency room encounters compared to people in the Low Treatment Need group (see Table 3).
Discussion Research demonstrates that multiple risk factors are the norm among U.S. adults (Coups et al., 2004; Fine et al., 2004; Pronk et al., 2004) and contribute to mortality (McGinnis & Foege, 1993). Given healthcare’s proliferating use of EMRs and the use of standard health screens, EMRs can be easily utilized to provide a simple way to identify patients who have multiple, modifiable health risk factors. This study aimed to advance the initial findings of Funderburk et al. (2008) by examining the relationship between classification in one of the three derived groups of patients and healthcare utilization as well as negative health-outcomes during a follow-up period of about 1 year. Results showed that individuals assigned to the High Treatment Need group had a greater number of diagnoses than individuals in the other groups, supporting the hypothesis that these individuals had a greater number of medical and psychiatric diagnoses within the follow-up time period compared to the other two groups. Also, individuals in the High Treatment Need group had a higher number of visits to primary care and the highest likelihood of having encounters at behavioral health clinics and the emergency room in the subsequent follow-up period compared to the other two groups. Due to the temporal
– 0.91 (0.75–1.11) 2.96 (2.29–3.83)***
– – 3.25 (2.44–4.30)***
ordering of the data, these results provide preliminary evidence of the predictive value of the risk factor profiles identified by Funderburk et al. (2008) in identifying which groups of patients were more likely to be diagnosed in the future with more medical and mental health diagnoses, and more likely to utilize their healthcare system. These findings are not surprising because the factors that were used to create the High Treatment Need group depended on patients screening positive on several health screens designed to identify individuals who are either currently experiencing or have a high likelihood of developing a disease/disorder. These data provide information to the healthcare system as to how often and where these patients are using services. Such information can be used to take the next step in identifying whether there are ways to help reduce the negative health-related consequences that these individuals tend to experience. Even though the results were generally consistent with our hypotheses, we did not find support for our prediction that individuals in the Moderate Treatment Need group would have a significantly greater number of diagnoses than individuals in the Low Treatment Need group. In this regard, it is possible that the follow-up period may not have been a sufficient amount of time for the development and detection of the detrimental effects of smoking and alcohol misuse. If that were the case, then the Low Treatment Need and Moderate Treatment Need groups would not appear as significantly different as hypothesized. Another potential reason for the result of no differences between the Low and Moderate Treatment Need groups is the similar long-term effects smoking and risky alcohol use can have on the cardiovascular system, such as high blood pressure. Research has shown that heavy alcohol use and smoking are independently associated with heart disease, hypertension, and other cardiovascular diseases/events (Dhaliwal & Welborn, 2009; Fuchs, Chambless, Whelton, Nieto, & Heiss, 2001; Rehm, Gmel, Sempos, & Trevisan,
123
16
2002; Rehm et al., 2003; U.S. Department of Health and Human Services, 1989). Because the initial LCAs identified a substantial portion of individuals in the Low Treatment Need group as screening positive for high blood pressure, it may be that these shared outcomes resulted in the Low Treatment Need and Moderate Treatment Need groups having fewer differences than expected. This study had limitations that are important to consider in interpreting its findings. First, participants were a random sample of Veteran primary care patients from Upstate New York, so that generalizability of the results is limited accordingly. This study’s findings also may not be applicable to patients seen in non-VA settings, because the VA typically treats an older population (Office of Policy and Planning, 2009) that tends to have poorer health status, more medical conditions, and a higher rate of medical care usage compared to the general population (Agha, Lofgren, VanRuiswyk, & Layde, 2000). In addition, there are mandates at the primary care provider level within the Veterans Health Administration for standard screening and specific protocols for potential treatments that may not match what occurs at primary care clinics outside of the VA. For example, Veterans screening positive for being overweight/obese are at a minimum offered the option of a national weight management program at the VA (Kisinger et al., 2009). Additionally, the data were limited by the degree of accuracy of patients’ EMRs. Some patients may have had incomplete records for a number of reasons, such as medical diagnoses and/or clinical visits not being properly documented, or alternative primary care, emergency room, or behavioral health services outside of the VA being utilized but not documented in patients’ VA medical records. As a result of these pitfalls in collecting data from medical records, we may not have had the most comprehensive scope of healthcare utilization or diagnoses, possibly affecting our findings. Third, it may have been the case for some patients that they already had the disorder/disease at the time of the initial screening, but a diagnosis was not entered into their EMRs until a later date. If that were true, then these patients would have had positive scores at the initial screening time (i.e., Time 1) and a subsequent diagnosis assigned during the follow-up (i.e., Time 2), leading to the inaccurate conclusion that these patients developed the disease in question during the follow-up period. Fourth, given that some disorders/diseases take longer to develop or to be detected than others, future research should incorporate a longer follow-up period. Fifth, because of the timeframe used for the follow-up period, it may have been that many patients did not need to utilize the behavioral health or emergency room services. The result was a predominance of zeros as a score on variables representing such encounters. Because of the
123
J Clin Psychol Med Settings (2014) 21:10–18
consequent non-normal distributions, logistic regression analyses were used to test part of our first main hypothesis instead of linear regression analyses. Future research should focus on not only replicating the findings that individuals classified to the High Treatment Need group have a greater likelihood of behavior health and emergency room encounters, but also to determine if these individuals actually have significantly more encounters than individuals in the other classes. Sixth, we did not obtain information from the EMRs regarding the specific interventions given to the patients in each group. The VA offers specific programs that target a single risk factor (e.g., smoking cessation program). However, each specific VA has the ability to modify that intervention or offer additional interventions that are unknown to the authors. This information will be necessary as we continue to examine these three groups of patients to determine how and in what context the healthcare system may best serve them. The results of this study underscore the usefulness of EMR data and the risk factor profiles that Funderburk et al. (2008) described in the identification of patients in need of early interventions to mitigate the development of future medical and/or mental health disorders due to modifiable multiple health risk factors. Our findings further highlight the need to examine the interventions that these groups of patients currently receive in behavioral health, primary care, and the emergency department and to evaluate whether existing or new interventions can simultaneously target the cluster of multiple health risk factors that appear to be present in this healthcare system. Because many chronic illnesses share common pre-morbid risk factors, interventions targeting these common risk factors in an integrated way may be more effective in reducing the risk of developing several chronic illnesses. As healthcare moves towards incorporating medical homes and using more team-based approaches to care, it is likely that interventions utilizing such teams, such as group medical visits, would be effective. Considering that primary care is the entry point into the healthcare system for most Americans (Institute of Medicine, 1990), it is an ideal setting in which to (a) ascertain which patients may be in need of a multiple risk intervention, and (b) administer treatment. By identifying patients with a particular risk profile through use of EMR data, it is possible that healthcare systems can evaluate whether tailoring interventions/treatments to best suit these patients in an effort to help further prevent disease progression or development may be more effective than the traditional intervention/treatment approach of addressing each health behavior separately. Acknowledgments This project was supported by funding from the Center for Integrated Healthcare, Syracuse Veterans Affairs Medical
J Clin Psychol Med Settings (2014) 21:10–18 Center, Syracuse, NY. All the work for this project was done at the Syracuse Veterans Affairs Medical Center, 800 Irving Ave., Syracuse, NY 13210. The views expressed herein are those of the authors and do not necessarily reflect those of the Department of Veterans Affairs or other departments of the U.S. government. Conflict of interest Jennifer S. Funderburk, Stephen A. Maisto, Allison K. Labbe declare that they have no conflict of interest. Ethical standard All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1975, as revised in 2000. A waiver of informed consent was obtained for all data collected from the electronic medical record.
References Agha, Z., Lofgren, R. P., VanRuiswyk, J. V., & Layde, P. M. (2000). Are patients at Veterans Affairs Medical Centers sicker? A comparative analysis of health status and medical resource use. Journal of the American Medical Association, 160, 3252–3257. An, L. C., Zhu, S. H., Nelson, D. B., Arikian, N. J., Nugent, S., Partin, M. R., et al. (2006). Benefits of telephone care over primary care for smoking cessation: A randomized trial. Archives of Internal Medicine, 166, 536–542. Berrigan, D., Dodd, K., Troiano, R. P., Krebs-Smith, S. M., & Barbash, R. B. (2003). Patterns of health behavior in U.S. adults. Preventive Medicine, 36, 615–623. Bliese, P. D., Wright, K. M., Adler, A. B., Cabrera, O., Castro, C. A., & Hoge, C. W. (2008). Validating the primary care posttraumatic stress disorder screen and the posttraumatic stress disorder checklist with soldiers returning from combat. Journal of Consulting and Clinical Psychology, 76, 272–281. Bush, K., Kivlahan, D. R., McDonell, M. B., Fihn, S. D., & Bradley, K. A. (1998). The AUDIT alcohol consumption questions (AUDIT-C): An effective brief screening test for problem drinking. Archives of Internal Medicine, 158, 1789–1795. Calfas, K. J., Long, B. J., Sallis, J. F., Wooten, W. J., Pratt, M., & Patrick, K. (1996). A controlled trial of physician counseling to promote the adoption of physical activity. Preventive Medicine, 25, 225–233. Coups, E. J., Gaba, A., & Orleans, T. (2004). Physician screening for multiple behavioral health risk factors. American Journal of Preventive Medicine, 27, 34–41. Critchley, J. A., & Capewell, S. (2003). Mortality risk reduction associated with smoking cessation in patients with coronary heart disease: A systematic review. Journal of the American Medical Association, 290, 86–97. Dhaliwal, S. S., & Welborn, T. A. (2009). Central obesity and cigarette smoking are key determinants of cardiovascular disease deaths in Australia: A public health perspective. Preventive Medicine, 49, 153–157. Ely, A. C., Banitt, A., Befort, C., Hou, Q., Rhode, P. C., Grund, C., … Ellerbeck, E. (2008). Kansas primary care weighs in: A pilot randomized trial of a chronic care model program for obesity in 3 rural Kansas primary care practices. Journal of Rural Health, 24, 125–132. Fine, L. J., Philogene, G. S., Gramling, R., Coups, E. J., & Sinha, S. (2004). Prevalence of multiple chronic disease risk factors: 2001 National Health Interview survey. American Journal of Preventive Medicine, 27, 14–18. Fuchs, F. D., Chambless, L. E., Whelton, P. K., Nieto, F. J., & Heiss, G. (2001). Alcohol consumption and the incidence of
17 hypertension: The Atherosclerosis Risk in Communities study. Hypertension, 37, 1242–1250. Funderburk, J. S., Maisto, S. A., Sugarman, D. E., & Wade, M. (2008). The covariation of multiple risk factors in primary care: A latent class analysis. Journal of Behavioral Medicine, 31, 525–535. Goldberg, D. P., Gater, R., Sartorius, N., Ustun, T. B., Piccinelli, M., Gureje, O., et al. (1997). The validity of two versions of the GHQ in the WHO study of mental illness in general health care. Psychological Medicine, 27, 191–197. Goldberg, D. P., & Williams, P. (1978). A user’s guide to the General Health Questionnaire. Windsor, UK: NFER-Nelson. Goldstein, M. G., Whitlock, E. P., & DePue, J. (2004). Multiple behavioral risk factor interventions in primary care: Summary of research evidence. American Journal of Preventive Medicine, 27, 61–79. Haapanen, N., Miilunpalo, S., Vuori, I., Oja, P., & Pasanen, M. (1997). Association of leisure time physical activity with the risk of coronary heart disease, hypertension, and diabetes in middleaged men and women. International Journal of Epidemiology, 26, 739–747. Hyman, D. J., Pavlik, V. N., Taylor, W. C., Goodrick, G. K., & Moye, L. (2007). Simultaneous vs. sequential counseling for multiple behavior change. Archives of Internal Medicine, 167, 1152–1158. Institute for Clinical Systems Improvement. (2006). Prevention and management of obesity (mature adolescents and adults). Bloomington, MN: Institute for Clinical Systems Improvement. Institute of Medicine. (1990). Alcohol problems: Broadening the base of treatment. Washington, DC: National Academy Press. Jonas, B. S., Franks, P., & Ingram, D. D. (1997). Are symptoms of anxiety and depression risk factors for hypertension?: Longitudinal evidence from the National Health and Nutrition Examination survey I epidemiologic follow-up study. Archives of Family Medicine, 6, 43–49. Katon, W., Fan, M., Unu¨tzer, J., Taylor, J., Pincus, H., & Schoenbaum, M. (2008). Depression and diabetes: A potentially lethal combination. Journal of General Internal Medicine, 23, 1571–1575. Kisinger, L. S., Jones, K. R., Kahwati, L., Harvey, R., Burdick, M., Zele, V., et al. (2009). Design and dissemination of the MOVE! Weight Management Program for Veterans. Preventing Chronic Disease, 6, 1–6. Klatsky, A. L. (1996). Alcohol and hypertension. Clinica Chimica Acta, 246, 91–105. Kroenke, K., Spitzer, R. L., & Williams, J. B. (2003). The PHQ-2. Medical Care, 41, 1284–1294. Lloyd-Jones, D., Adams, R., Carnethon, M., De Simone, G., Ferguson, T. B., Flegal, K., et al. (2009). A report from the American Heart Association Statistics Subcommittee and Stroke Statistics Subcommittee. Circulation, 119, 480–486. McGinnis, J. M., & Foege, W. H. (1993). Actual causes of death in the United States. Journal of the American Medical Association, 270, 2207–2212. National Institute on Drug Abuse. (2006). Tobacco addiction (No. NIH Publication No. 06-4342). Rockville, MD: National Institute on Drug Abuse. Ockene, J. K., Adams, A., Hurley, T. G., Wheeler, E. V., & Hebert, J. R. (1999). Brief physician and nurse practitioner-delivered counseling for high-risk drinkers: Does it work? Archives of Internal Medicine, 159, 2198–2205. Office of Policy and Planning. National Center for Veterans Analysis and Statistics. (2009). Analysis of unique veterans utilization of VA benefits & services. http://www.va.gov/vetdata/Report.asp. Accessed 12 July 2012.
123
18 Prins, A., Ouimette, P., Kimerling, R., Cameron, R. P., Hugelshofer, D. S., Shaw-Hegwer, J., … Sheikh, J. I. (2003). The primary care PTSD screen (PC-PTSD): Development and operating characteristics. Primary Care Psychiatry, 9, 9–14. Prochaska, J. O. (2008). Multiple health behavior research represents the future of preventive medicine. Preventive Medicine, 46, 281–285. Pronk, N. P., Anderson, L. H., Crain, A. L., Martinson, B. C., O’Connor, P. J., Sherwood, N. E., … Whitebird, R.R. (2004). Meeting recommendations for multiple healthy lifestyle factors: Prevalence, clustering, and predictors among adolescent, adult, and senior health plan members. American Journal of Preventive Medicine, 27, 25–33. Rehm, J., Gmel, G., Sempos, C. T., & Trevisan, M. (2002). Alcoholrelated morbidity and mortality. Alcohol Research & Health, 27, 39–51. Rehm, J., Room, R., Graham, K., Monteiro, M., Gmel, G., & Sempos, C. T. (2003). The relationship of average volume of alcohol consumption and patterns of drinking to burden of disease: An overview. Addiction, 98, 1209–1228. Richardson, L. K., Egede, L. E., Mueller, M., Echols, C. L., & Gebregziabher, M. (2008). Longitudinal effects of depression on
123
J Clin Psychol Med Settings (2014) 21:10–18 glycemic control in veterans with Type 2 diabetes. General Hospital Psychiatry, 30, 509–514. Smit, F., Beekman, A., Cuijpers, P., de Graaf, R., & Vollebergh, W. (2004). Selecting key variables for depression preventions: Results from a population-based prospective epidemiological study. Journal of Affective Disorders, 81, 241–249. Strik, J. J. M. H., Denollet, J., Lousberg, R., & Honig, A. (2003). Comparing symptoms of depression and anxiety as predictors of cardiac events and increased health care consumption after myocardial infarction. Journal of the American College of Cardiology, 42, 1801–1807. U.S. Department of Health and Human Services. (1989). Reducing the health consequences of smoking: 25 years of progress: A report of the surgeon general (DHHS Publication No. CDC 89-8411). Rockville, MD: Office on Smoking and Health. U.S. Department of Health and Human Services. (2003). The seventh report of the Joint National Committee on prevention, detection, evaluation, and treatment of high blood pressure (No. NIH Publication No. 03-5233). World Health Organization Brief Intervention Study Group. (1996). A cross-national trial of brief interventions with heavy drinkers. American Journal of Public Health, 86, 948–955.
