Currently Viewing:
Currently Reading
Patterns and Predictors of Medication Initiation in Diabetic Peripheral Neuropathy
Chris M. Kozma, PhD; Terra l. Slaton, MS; Wing Chow, PharmD, MPH; and Marcia F. T. Rupnow, PhD
Budgetary Impact Analysis of Denosumab in a US Health Plan
Anju Parthan, PhD; Nicholas P. Emptage, MAE; Douglas C. A. Taylor, MBA; Hema N. Viswanathan, PhD; Nicole Yurgin, PhD; Bradley Stolshek, PharmD; Karen M. Clements, ScD; Charles Y. Tao, MA; and Milton C

Patterns and Predictors of Medication Initiation in Diabetic Peripheral Neuropathy

Chris M. Kozma, PhD; Terra l. Slaton, MS; Wing Chow, PharmD, MPH; and Marcia F. T. Rupnow, PhD
This analysis of more than 50,000 patients examined medication use (anticonvulsants, opioids, and antidepressants) for diabetic peripheral neuropathy.
Objectives: To investigate patterns and predictors of medication initiation and type of medication initiated in patients with newly diagnosed diabetic peripheral neuropathy (DPN).

Study Design: Retrospective analysis of a nationwide administrative claims database.

Methods: Adults with a new principal DPN diagnosis were categorized as initially untreated if they had no newly initiated DPNrelated medication within the fi rst 14 days after diagnosis. Bivariate logistic regression evaluated predictors for newly initiating medication versus being initially untreated. Multinomial logistic regression evaluated predictors for each medication category (antidepressants, anticonvulsants, or multiple medications versus opioids).

Results: Of 53,753 patients with DPN, 46,335 (86%) were initially untreated and 7418 (14%) newly initiated an anticonvulsant (48%), opioid (28%), antidepressant (18%), or multiple medications (6%). Having a preindex claim for an anticonvulsant, antidepressant, or weak opioid predicted newly initiating another DPN-related medication. A weak opioid preindex predicted switching to another medication category after diagnosis and a strong opioid preindex predicted switching to an antidepressant or multiple medications. A preindex anticonvulsant predicted switching to an opioid. A preindex antidepressant was not predictive of switching. Of the patients who newly initiated medication, 28% had received an opioid preindex and 43% to 74% received an opioid between 14 and 180 days postindex.

Conclusions: Prior medication use is the most signifi cant predictor of whether patients initiate a new medication at the time of a new DPN diagnosis, as well as what type of medication they receive. Opioid use is common before and after starting a new medication for DPN.

Am J Pharm Benefits. 2013;5(5):e111-e121
Improved understanding of pain progression and consideration of what constitutes optimal pain management are needed in planning for diabetic peripheral neuropathy (DPN) medication pharmacy benefits.

  •  At DPN diagnosis, 14% of patients initiated a DPN-related medication, mostly an anticonvulsant or antidepressant (66%), followed by opioids (28%) and combination treatment (6%).

  • Having a DPN-related medication claim before DPN diagnosis predicted initiating another DPN-related medication at diagnosis, illustrating the diffi ulties in finding optimal regimens.

  • Many patients received opioids before or after their diagnosis.
Of the more than 25 million Americans with diabetes, 1,2 up to 25% will develop painful diabetic peripheral neuropathy  (DPN).3,4 Patients with DPN are generally in poorer health and incur higher healthcare resource utilization and costs than similar patients without DPN.5 Moreover, these patients are more likely to have greater numbers of comorbidities including musculoskeletal and other neuropathic pain–related conditions,6,7 increasing the likelihood that they will be prescribed pain medications. Given the clinical and economic burden associated with painful DPN, deeper understanding of factors that lead to medication initiation and medication choice is necessary for optimization of patient outcomes.

Painful DPN is a challenging chronic condition to manage and often requires a multimodal approach, in parallel or sequentially. Guidelines recommend antidepressants or anticonvulsants as fi rst-line treatments in the management of painful DPN.8-10 Of the antidepressants and anticonvulsants available in the United States, only duloxetine (an antidepressant) and pregabalin (an anticonvulsant) are indicated for the management of painful DPN.11,12 However, a metaanalysis of the available data suggested that other antidepressants (tricyclic antidepressants, selective serotonin reuptake inhibitors) and anticonvulsants (gabapentin, sodium-channel blockers, membrane-stabilizing agents) also reduce chronic neuropathic pain more effectively than placebo.13

When antidepressants or anticonvulsants provide insufficient pain relief, guidelines for the management of painful DPN  recommend opioid therapy.8-10 Limited literature examining treatment patterns within patients with painful DPN has suggested that opioids are a commonly used class of medications,7,13 but the reasons and timing for introducing an opioid regimen are not well documented. Given the chronic nature of DPN, treatment with a long-acting opioid (sustained-release, extended-release, or controlled-release formulations) is a recommended component of multimodal management strategies.14 Currently, only 1 opioid is indicated for the management of painful DPN in the United States (tapentadol extended release).15  Nonetheless, patients with DPN are up to 10 times more likely to receive a short-acting opioid than a long-acting opioid.2,7,16-20 To our knowledge, predictors for the use of long-acting or short-acting opioids in the management of painful DPN have not been explored previously.

The primary objective of this analysis was to investigate predictors of medication initiation and type of medication initiated in newly diagnosed DPN patients.


Data Source

The data source for this analysis was employer-based claims from the MarketScan Research Database.21 The annual medical databases include private-sector health data from approximately 100 payers and more than 500 million claim records. The database represents the medical experience of insured employees and their dependents for active employees, early retirees, those who continue on COBRA, and Medicare-eligible retirees with employer-provided Medicare Supplemental plans. The database includes integrated patient-level data for inpatient, outpatient, and drug claims, as well as insurance eligibility. The database is de-identified and compliant with the Health Insurance Portability and Accountability Act of 1996. As such, no institutional review board approval was necessary.

Study Design

This retrospective analysis included claims from January 1, 2005, through December 31, 2009 (Figure 1). A diagnosis of DPN was identified from International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) codes 250.6X (diabetes with neurologic manifestations) or 357.2X (polyneuropathy in diabetes). The index date was the date of the first principal diagnosis of DPN in the identifi cation window (January 1, 2006, to June 30, 2009), which was selected to allow for analysis of claims data for 360 days preindex through 180 days postindex. To be eligible for inclusion, patients were required to be at least 18 years of age at the index date and continuously eligible for insurance plan and pharmacy coverage throughout their preindex and postindex periods. Patients were excluded if they had a diagnosis of DPN in the 360-day preindex period (ie, only “newly” diagnosed patients were analyzed).

Patients were divided into 2 cohorts: the newly initiated medication group and the initially untreated group. Because prescriptions were not linked to a specifi c diagnosis, rules were defi ned to identify patients who received medications for the DPN diagnoses. The goal of the rules was to create a high degree of confidence that the new medication therapy was related  to the DPN diagnosis. The newly initiated medication group included patients who received a new prescription for a DPN-related medication (Appendix) within 14 days postindex. The 14-day window was selected to identify medicationsthat were likely to have been initiated for the new DPN diagnosis; it was long enough to allow patients up to 2 weeks to fill their new prescription after the DPN diagnosis, but short enough to exclude DPN-related medications that were prescribed for subsequent diagnoses. Patients in the newly initiated medication group could have used other medications (including another medication from the same category) previously, but they were excluded from the analysis if they had another claim in the 360-day preindex period for the same medication (eg, generic level) that they received within 14 days postindex.

To increase certainty that patients were using medications related to DPN and not for other purposes, patients in the newly initiated medication group were excluded from the analysis if they had any of the following diagnoses between the DPN diagnosis and the newly initiated medication: a newly initiated antidepressant and a mental health diagnosis (ICD-9-CM codes 290.xx-319.xx); a newly initiated anticonvulsant and a diagnosis of epilepsy and recurrent seizures or convulsions (ICD-9-CM codes 345.xx or 780.3); a newly initiated antidementia agent and a diagnosis of a mental disorder or disease of the nervous and sense organs (ICD-9-CM codes 290.xx-319. xx or 320.xx-389.xx); or a newly initiated antihypertensive or antiarrhythmic agent and disease of the circulatory system (ICD-9-CM codes 309.xx-459.xx).

The newly initiated medication group was divided into subgroups by the category of medication received: antidepressant, anticonvulsant, opioid, or multiple medications. The analysis was originally designed to include antiarrhythmics, antidementia agents, antihypertensives, or topical anesthetics, but these categories occurred too infrequently (<0.23% of the total population) for meaningful interpretation; therefore, patients who were newly initiated on these medications were excluded.

The initially untreated group included patients who did not receive a new DPN-related medication within the first 14 days postindex. Patients were excluded from the initially untreated group if they had received a DPN-related medication during the 60 days leading up to the index date or if they received their first DPN-related medication between postindex days 14 and 31. The purpose of these exclusion criteria was to reduce ambiguity by removing patients whose medications might have been related to the new DPN diagnosis but did not satisfy the criteria for inclusion in the newly initiated medication group.

Patients with any of the following were excluded from the study: a documented diagnosis of malignant neoplasm (ICD-9-CM codes 140.XX-172.XX, 174.xx- 209.3X, or 209.7x) during the study period (2005-2009); pregnancy (ICD-9-CM codes 630.XX‑679.XX, or V22.XX); during the preindex or postindex period; negative, zero, or missing days of supply or quantity on DPN-related medication claims; or long-term inpatient care, custodial care, or hospice (place of service codes 27, 33, and 34, respectively) during the study period.

Statistical Analysis

All analyses were performed using SAS for Windows version 9.2 (SAS Institute Inc, Cary, North Carolina). For descriptive analyses, categorical variables were summarized using frequencies and percentages, and continuous variables were summarized with means and standard deviations. Two independent logistic regressions were conducted in which the possible predictors were age, sex, wage status, Medicare indicator, resource use, comorbidities, and drug use variables. The World Health Organization (WHO) analgesic ladder was used to categorize drug use variables as nonopioids (WHO-1), weak opioids(WHO-2), and strong opioids (WHO-3). The bivariate logistic regression predicted whether patients newly initiated medication or were initially untreated. Odds ratios (ORs) and 95% confidence intervals (CIs) were determined for each predictor, and goodness of fit was examined with a Hosmer-Lemeshow test. A multinomial logistic regression predicted which DPN-related medication (anticonvulsant, opioid, antidepressant, or multiple medications) was newly initiated, with opioids as the reference category. An exploratory analysis used logistic regression to evaluate predictors for long-acting opioids versus short-acting opioids among patients with newly initiated opioids. Another exploratory analysis examined the frequency and percentage of patients (initially untreated or by newly initiated medication) with subsequent claims for DPN-related medications 14 to 180 days postindex. 


Analyzed Patients

Of the 290,449 patients with any diagnosis code for DPN in the study period, a total of 53,753 had a new principal DPN diagnosis in the identification period and satisfied all other study inclusion criteria (Figure 2). There were 46,335 (86%) initially untreated patients with no newly initiated medication and no prior exposure to DPN-related medications within 60 days before the DPN diagnosis. Of the 7418 patients who newly initiated medication, 3589 (48%) received an anticonvulsant, 2068 (28%) received an opioid, 1316 (18%) received an antidepressant, and 445 (6%) received multiple medications.

Patient baseline characteristics are provided in Table 1. During the 360-day preindex period, the percentage of patients who received an anticonvulsant, opioid, or antidepressant (other than their index DPN-related medication) was 27%, 53%, and 28%, respectively, in the newly initiated medication group, and 3%, 22%, and 3%, respectively, in the initially untreated group. Patients in the newly initiated medication group had higher rates than initially untreated patients for preindex emergency departmen tvisits (35% vs 24%) and inpatient hospitalizations (23% vs 15%). The newly initiated medication group also incurred higher costs (mean = $16,699; standard deviation [SD] = $31,751) during the preindex period than the initially untreated group (mean = $11,237; SD = $24,098).

Predictors for Newly Initiating a DPN Medication

Copyright AJMC 2006-2019 Clinical Care Targeted Communications Group, LLC. All Rights Reserved.
Welcome the the new and improved, the premier managed market network. Tell us about yourself so that we can serve you better.
Sign Up