Hospital admissions, emergency department utilisation and patient activation for self-management among people with diabetes

Abstract

Aims

To assess the relationship between patient activation for self-management and admissions to hospital or attendances at emergency departments among people with diabetes, after controlling for other known associations.

Methods

Patients were randomly selected from Australia's National Diabetes Services Scheme and invited to participate in the Living with Diabetes Study, which is a longitudinal survey providing a comprehensive examination of health care utilisation, well-being and disease progression. Data was collected for 3951 participants.

Results

Outcome events were defined as 1 or more hospitalization and 1 or more visits to an emergency department in the preceding 12 months. Logistic regression analyses showed six variables remained significantly associated with both outcomes: age, income, disease duration and severity, current depression and PAM stage. Patients at PAM stage 1 were 1.4 times more likely to be hospitalised (p = 0.023) and 1.3 times more likely to have visited emergency (p = 0.049) compared to those at stage 4.

Conclusions

Low levels of activation are associated with higher utilisation of hospital resources even after controlling for relevant factors such as disease severity and co-morbid depression. Most will be gained by moving patients from PAM stage 1 to a higher level of activation.

Introduction

The prevalence of chronic disease, and particularly diabetes, is increasing worldwide, thereby intensifying the demand for health services, including inpatient care [1]. The economic consequences of both the high use of medical services for diabetes treatment, and the growing number of patients with the condition, will be substantial. Patients with diabetes are at higher risk of hospitalization, longer stay in hospital, and greater total inpatient costs than the general population [2], [3], [4], [5]. In Australia, where this research is based, hospital admission for any diagnosis of diabetes mellitus increased by 35% between 2000–2001 and 2004–2005 [6]. Elsewhere, it has been reported that patients with diabetes are hospitalised up to three times more often than those without the condition, and diabetes patients are likely to stay in hospital 30% longer [4], [5]. Low income backgrounds, longer disease duration, disease severity and co-morbid depression are all important determinants of hospital resource utilisation for people with diabetes [7], [8], [9], [10].

The hospitalization of people with diabetes is commonly precipitated by conditions like heart disease, which may or may not be the result of their diabetes [11]. Frequently in these situations, the management of diabetes becomes secondary to that of the primary diagnosis. Poor glycaemic control is therefore common among hospital inpatients with diabetes, particularly those treated with insulin. Both length of stay in hospital and the number of readmissions may increase when symptoms of hypoglycaemia develop [12]. Likewise, improved glycaemic control is associated with fewer inpatient admissions and fewer emergency department visits [13], [14], [15]. Furthermore, intensive glycaemic control has been shown not to result in increased emergency department visits for hypoglycaemia [16].

Patient self-management of diabetes has been widely recognised as an important contributor to improved health outcomes [17]. Improved self-management of diabetes may prevent short-term complications like hypoglycaemic episodes, infections, and electrolyte disturbances; as well as decrease the risk of long-term complications of diabetes [14], [18]. However, the evidence supporting the value of self-management education for diabetes is mixed [19]. In one systematic review, no studies demonstrated that self-management training improved cardiovascular disease outcomes [19] and of those studies that examined health service utilisation, most failed to demonstrate improvements [19]. Furthermore, lay-led self-management education programs did not reduce time spent in hospital by people with chronic health conditions [20].

It is likely that the success of self-management education programs is dependent on the patient. Hibbard and colleagues developed the Patient Activation Measure (PAM), which assesses a person's beliefs about, motivation for, and action for self-care [21]. Diabetes patients with higher levels of activation for self-management enjoy better health outcomes than those with lower scores [22], [23], [24]. The levels of patient activation are important because engaged, informed, confident, and skilled patients are more likely to perform activities that will promote their own health, and are more likely to have their health care needs met [25].

Retrospective analysis of secondary data has revealed that PAM scores are predictive of all-cause hospitalizations among patients with diabetes [24]. In this paper we use cross-sectional data from a large cohort of patients diagnosed with either type 1 or type 2 diabetes to further explore the relationship between patient activation for self-management and two outcomes of interest: (i) hospitalization for diabetes-related conditions and (ii) emergency department presentations for diabetic patients. The study is of sufficient magnitude to control for known risk factors such as disease duration and disease severity, and in multivariate models.