Measuring Total Healthcare Utilization Among Intimate Partner Violence Survivors In Primary Care

1. INTRODUCTION

In the United States, a disproportionate share of health care spending is attributed to a relatively small group of patients. Thus, as national health care costs continue to rise, health care utilization is a critical research focus. The Agency for Healthcare Research and Quality estimates the top 5% of patients account for approximately 50% of U.S. health care expenditures, while the top 1% of patients account for 21% of healthcare costs (Cohen and Yu 2012). Health care utilization is most often analyzed within the economic context of volumes of services rendered (Da Silva et al. 2011), with the term ‘high-utilizer’ applied to patients who accumulate large numbers of emergency department visits or hospital admissions (Mann 2013). However, health care utilization should be expanded to include a wide range of ambulatory services and non-face-to-face episodes of care, whereby a patient may interact with the health system outside of a traditional office visit. Although it is not customary to bill for non-face-to-face episodes under a traditional fee-for-service model of care, these encounters may require significant provider time or clinic resources (Christino et al. 2013). Non-face-to-face encounters include telephone calls, prescription refills, email messages, or review of reports or records (Baron 2010). The authors of one observational analysis of ambulatory activities across four specialties found that physicians spent only 33.1% of their time on direct clinical face time with patients and staff (Sinsky et al. 2016). The consequences of this allocation of effort ranged from lost revenue (an estimated $4–6 for every minute spent), to poorer patient-physician interactions, worse patient health outcomes, and greater physician burnout (Gottschalk and Flocke 2005; Shanafelt et al. 2012).

A potential solution to the utilization and payment incentive crisis requires us to identify high-utilizers of care in order to further target gaps in our care delivery system and deliver interventions to those at highest risk. High-utilizers often lack access to regular primary care, are of lower socioeconomic status, have complex medical conditions, and face challenging psychosocial factors, such as trauma, homelessness, substance abuse or mental illness (Bell et al. 2017; Mann 2013). Additionally, females account for a larger overall burden of health care costs compared to males, owing in part to reproductive health needs and longer lifespans (Owens 2008). One additional psychosocial factor associated with more frequent utilization of health care services is intimate partner violence (IPV). IPV, defined as physical violence, sexual violence, stalking and/or psychological aggression by a current or former intimate partner, affects an estimated one in three females and one in four males in their lifetime and leads to numerous adverse physical, mental and behavioral health outcomes (Breiding et al. 2014; Breiding, Chen, and Black 2014; Breiding, Black, and Ryan 2008). IPV is associated with increased utilization of emergency care and other services, even after violence has stopped (Fishman et al. 2010; Jones et al. 2006). Females who experienced IPV in the past year had health care costs that were 122% higher than females without a history of IPV, and 60% higher than females who experienced IPV prior to the past year (Bonomi et al. 2009; Rivara et al 2007). Using data from the Medical Expenditure Panel Survey and the National Violence Against Women Survey, the estimated costs of IPV at the U.S. population level are $2.6 billion per year (Brown, Finkelstein, and Mercy 2008; Max et al. 2004). Due to the prevalence of IPV victimization and its exorbitant costs, understanding how IPV is associated with increased health care utilization is the first step in disrupting this association.

In this study, we built upon existing IPV and utilization literature by expanding the scope of health care utilization examined. To date, there are no published findings specific to IPV or chronic disease where researchers examined non-face-to-face encounters when assessing utilization— despite the growing frequency, associated costs, and professional dissatisfaction with these encounters (Blumenthal et al. 2017; Heincelman et al. 2016; Meijs, et al. 2014; Robinson et al. 2014; Seibert et al. 2008). To fill this gap, we developed a novel health care utilization scoring tool—which encompasses both face-to-face encounters and non-face-face encounters—to assess whether patients who experience IPV have a higher utilization of health care services compared to those who do not when controlling for relevant confounders. Through this study, we sought to shed light on how IPV victimization is associated with differences in utilization for females and males; and consequently, how population-specific interventions may improve health care expenditures, outcomes and provider satisfaction, satisfying the ‘quadruple aim’ of health care (Bodenheimer and Sinsky 2014).

2. METHODS

3. RESULTS

During the time period of interest, 31,305 unique patients were screened for IPV victimization (Table 3). The positive screening rate was 0.9% (n=280). Of the 280 patients who screened positive, 79.6% were female and 20.4% were male. The mean age of the sample was 45.2 years old and 49.0% of patients were non-Hispanic White, which is less than the census-level reported data for the metro region, suggesting that patients screened at the primary care clinics in this study may represent a more diverse segment of the population. The mean Charlson Comorbidity Index among patients who screened positive for IPV was 0.86 versus 0.87 for those who screened negative (p=0.97), with a combined Charlson Age Score of 1.50 and 1.64 (p=0.30), respectively. This indicated that most patients had one chronic health condition and that the two groups did not differ significantly with respect to comorbid illness.

Patients who experienced IPV had a significantly higher number of missed appointments compared to patients who did not experience IPV (3.0 versus 1.3 respectively, p<0.001). Similarly, patients who experienced IPV had more overall health care utilization. IPV-positive patients had an average of 7 more non-face-to-face encounters and 4 additional outpatient face-to-face encounters over the two-year analysis period than IPV-negative patients (p<0.001). In-patient encounters (hospital, emergency department, and surgeries) were not significantly different for the two groups (1.5 versus 1.8, p=0.15). Despite this, the overall weighted utilization score was 54 for IPV-positive patients and 40 for IPV-negative patients (range 2-369, p<0.001).

In the multivariate regression analyses (Table 4), we found total health care utilization was associated with a higher Charlson Comorbidity score (p<0.001) and older age (p<0.001). Patients who were not non-Hispanic White were less likely to utilize care than non-Hispanic Whites (p<0.001). The main effect of IPV was not significant (p=0.7); however, the interaction between IPV and sex was significant (p=0.015). Females who experienced IPV were significantly more likely to utilize health care resources than both females who did not experience IPV (p<0.001) and males, with or without IPV experiences (p=0.004 and p<0.001 respectively). IPV-positive males were not more likely to utilize health care resources than IPV-negative males.

When examining different types of health care utilization, similar patterns held true for both non-face-to-face encounters and ambulatory visits, with IPV-positive females utilizing more resources. However, for in-patient encounters (e.g., hospital admissions, emergency room visits and surgeries), we found that neither the main effect of IPV (p=0.16), nor the interaction between IPV and sex was significant (p=0.37). Being female, non-Hispanic White, of younger age, and having a higher Charlson Comorbidity score were significantly associated with greater numbers of hospital encounters. Being female, not identifying as non-Hispanic White, being younger, having a higher Charlson Comorbidity score, and experiencing IPV victimization (females and males) (p<0.001) were significantly associated with missed appointments.

In summary, female patients had significantly greater total health care utilization, non-face-to-face encounters, and missed appointments (p<0.001). The main effect of IPV was associated with increased rates of missed appointments (p<0.001) and the interaction between sex and IPV was associated with increased total utilization (p=0.015), as well as non-face-to-face and ambulatory visits for females (p=0.025 and p=0.015, respectively).

4. DISCUSSION

To our knowledge, we are the first to create a health care utilization tool that includes non-face-to-face encounters and stratifies encounters by acuity (and effectively cost) in order to create a summed score. We investigated both traditional predictors of utilization, including race/ethnicity, sex, age, and medical complexity, as well as novel variables such as IPV victimization. We also examined the effects of IPV for both males and females.

Overall, the results suggest that females who experience IPV contribute higher costs to the health care system via higher utilization of care and more missed appointments than females who do not report a history of IPV. Previous studies have identified higher health care costs (Bonomi et al. 2009; Rivara et al. 2007) as well as poor utilization of prenatal care among female IPV survivors (Cha and Masho 2014). This study expands upon this literature by further characterizing associations between IPV and various types of health care encounters, including ambulatory and non-face-to-face care. We also looked at associations between IPV, utilization and missed appointments across the complete spectrum of medical and surgical specialties, rather than focusing only on women’s health or mental health visits (Cha and Masho 2014; Illangasekare S et al. 2012).

Additionally, although we found that female but not male survivors of IPV utilized more health care services, both males and females had higher numbers of missed appointments compared to patients without a reported history of IPV. This suggests that the instability that characterizes some survivors’ lives similarly influences females’ and males’ abilities to keep scheduled appointments, although it does not necessarily drive both equally towards utilization of health resources. These associations contribute negatively to costs of care and health outcomes, and support a more expansive approach to IPV screening or identification than the current clinical recommendation which focuses only on women of reproductive age (Moyer et al 2013; IOM 2011).

Of note, the IPV victimization detection rate in our study was lower than previously reported estimates of prevalence (Breiding et al. 2014) and the IPV-positive cohort included a larger than anticipated proportion of males. It is important to be aware, however, that the lower detection rate was generated through the use of two safety-focused questions that are not IPV-specific nor are a valid IPV screen (Moyer 2013). The clinic system has since adopted a 4-item validated IPV screen for use with all adult patients. It would be valuable to replicate the current study to compare rates of IPV among patients, in addition to further testing the utilization score in an effort to ensure robust functioning.

5. CONCLUSION

Our results highlight an opportunity to intervene to impact both health outcomes and expenditures, particularly with respect to females and individuals who identify as racial and ethnic minorities who have experienced IPV and are utilizing more care. This is in addition to males and individuals who identify as sexual and gender minorities who may be under-recognized by health care professionals. The Institute of Medicine recommends including documentation of IPV and other social determinants of health in the electronic health record (Institute of Medicine 2014). Our study findings provide evidence to support a more coordinated and inclusive effort to identify and support patients with counseling, resources and referrals in our health care system (Institute of Medicine 2011). Additionally, while many health care organizations and payers are creating complex risk assessment models to evaluate costs and manage populations, this utilization measure has the benefit of being a comprehensive sum of health services utilization, which requires only EMR data and is replicable by scholars and clinicians.