National study of prescription poisoning with psychoactive and nonpsychoactive medications in Medicare/Medicaid dual enrollees age 65 or over.
Poisoning (Risk factors)
Psychotropic drugs (Health aspects)
Psychotropic drugs (Research)
Blackwell, Steven A.
Baugh, David K.
Ciborowski, Gary M.
Montgomery, Melissa A.
|Publication:||Name: Journal of Psychoactive Drugs Publisher: Taylor & Francis Ltd. Audience: Academic Format: Magazine/Journal Subject: Health Copyright: COPYRIGHT 2011 Taylor & Francis Ltd. ISSN: 0279-1072|
|Issue:||Date: Sept, 2011 Source Volume: 43 Source Issue: 3|
|Topic:||Event Code: 310 Science & research|
|Product:||Product Code: 2834250 Psychotherapeutic Preparations NAICS Code: 325412 Pharmaceutical Preparation Manufacturing SIC Code: 2834 Pharmaceutical preparations|
|Geographic:||Geographic Scope: United States Geographic Code: 1USA United States|
Abstract--The purpose of this study is to assess prescription
medication poisoning among psychoactive and nonpsychoactive medications
used by elderly (65 years or older) Medicare & Medicaid dual
enrollees as well as examine contextual components associated with
poisoning. Our primary research goal was to compare medication
poisonings among psychoactive medications to nonpsychoactive
medications. Our second research goal was to identify components
influencing medication poisonings and how they interrelate. The approach
used a cross-sectional retrospective review of calendar year 2003
Centers for Medicare & Medicaid Service's Medicaid Pharmacy
claims data for elderly dual enrollees. Poisonings were identified based
on ICD-9-CM categorizations. Poisonings associated with the psychoactive
medications were proportionally over twice as high as compared to
nonpsychoactive medications ( 14.3 per 100,000 enrollees and 6.6 per
100,000 enrollees, respectively). Additionally, the two contextual
components of (a) use of many drugs and (b) familiarity with the
medication have a direct, but competing impact on poisoning. The reasons
behind unintentional poisoning in the elderly have been somewhat a
mystery. This study is among the first to attempt to distinguish between
poisoning events associated with psychoactive medications versus
nonpsychoactive medicatious as well as assess the impact of differing
contextual components on medication poisoning.
Keywords--elderly, poisoning, safety
All prescription drugs--whether legally prescribed by a practitioner, purchased over the counter, or secured by other means such as provided from a friend or relative--have the potential to be misused. Taken in an excessive amount, even drugs normally considered safe can lead to adverse outcomes. Drug overdosing, either unintentional or intentional, is commonly referred to as poisoning (Camidge, Wood & Bateman 2003).
Almost all unintentional poisoning deaths are believed to be attributable to drugs--both prescription drugs and illegal drugs (Paluozzi, Bailesteros & Stevens 2006). A study using New Mexico data found not only that the trend in the number of unintentional prescription drug poisonings is rising, but that age is a contributing factor for prescription drug overdoses. (CDC 2010; Paluozzi & Annest 2007).
Prescription medication overdose has exploded in recent years. The Centers for Disease Control and Prevention (CDC) indicate that between 1999 and 2006, the number of poisoning deaths in the United States nearly doubled (CDC 2009). According to the CDC, the finding is largely due to overdose deaths involving prescription opioid painkiilers, a type of psychoactive medication.
In this study, we assess prescription medication poisoning among psychoactive and nonpsychoactive medications as well as examine contextual components associated with poisoning. We focus on the elderly (i.e., age 65 or older) population and use national enrollment data to conduct the study. Our primary research goal was to compare medication poisonings among psychoactive medications to nonpsychoactive medications. Individual psychoactive and nonpsychoactive medications as well as demographic factors that may be potential risk factors for medication poisonings were assessed. Our second research goal was to identify components influencing medication poisonings and how they interrelate.
Data and Analysis Plan
Our analysis was based on a cross-sectional study design. Our study population was the entire elderly dually eligible Medicare and Medicaid population aged 65 or over for the 2003 calendar year. The Center for Medicaid & Medicare Services' (CMS) Outpatient Standard Analytical file and Medicare Provider Analysis and Review file were used to identify medication poisonings resulting in emergency room visits during the 2003 calendar year. Admitting diagnoses pertaining to medication poisonings were identified on these claims data using the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes (National Center for Health Statistics 2010). The codes utilized indicate that the patient has been poisoned by a medication in a particular medication class as identified by the ER physician (i.e., physician-identified poisonings). The ICD-9-CM codes referring to the following categories were classified as psychoactive medication poisonings: analgesics, antipyretics and antirheumatics; anticonvulsants and anti-parkinsonism drugs; barbiturates and other sedatives; central nervous system (CNS) depressants and anesthetics; and tranquilizers, antidepressants, and antipsychotics. The ICD-9-CM codes referring to the following categories were classified as nonpsychoactive medication poisonings: anti-infectives; hormones and synthetic substitutes; primarily systemic agents; agents primarily affecting blood constituents; agents primarily affecting the cardiovascular system; agents primarily affecting the gastrointestinal system; and poisoning by water, mineral, and uric acid metabolism drugs.
Prescription medication fills, both original and refill, were captured beginning 90 days prior to calendar year 2003 and ending 90 days prior to December 31, 2003. Using the ICD-9-CM codes to identify that a medication poisoning occurred, we looked back 90 days prior to the ER visit to match the specific medication utilized by the patient to the physician-identified poisoning. Physician-identified poisoning patients who had the offending medication (i.e., as a prescription fill within 90 days of the admission) as identified by the ICD-9 poisoning code comprised our study population. Throughout this manuscript, we refer to beneficiaries identified in our study population as "PIPs" (physician-identified poisonings).
Data for prescription medication fills as well as age, gender and ethnicity, were obtained using the Medicaid Analytic Extract (MAX) files (CMS 2009). Prescription medication fills were captured tbr elderly dually eligible Medicare and Medicaid beneficiaries who would have attained age 65 in the 50 states and D.C. as of December, 2003 and consisted of community-dwelling residents only. Residents of long-term care facilities were excluded from the analysis. National Drug Code (NDC) data from the MAX files were combined with medication names from the Medi-Span therapeutic classification system to obtain the names and therapeutic categories of the medications we assessed (Wolters Kluwer Health 2010).
We assessed time to a poisoning event for each PIP patient based on the first recorded fill of the medication during this 90-day window. Following the removal of two beneficiaries from our study population who had more than one poisoning event, we performed individual-level analyses.
Our analysis plan was divided into two distinct phases. First, we identified PIPs and assessed the underlying use rates of the problematic medication (i.e., PIPs per 100,000 beneficiaries receiving the problematic medication). This allowed us to compare the proportion of PIPs associated with psychoactive medications to the proportion of PIPs associated with nonpsychoactive medications. For each proportion, we calculated the ratio of medication fills having a related PIP to total fills for the medication. We report our findings as population parameter assessments based on ICD-9-CM categories and beneficiary demographics. Second, we developed and tested a theoretical model in order to investigate direct-effect components that influence the timing of PIPs.
Patient demographics of PIP population. Descriptive analyses were initially performed to identify PIPs based on gender, age group, ethnicity, and census region. We then report PIPs by ICD-9-CM category including psychoactive versus non-psychoactive classification.
Theoretical model. We developed and tested a conceptual framework in order to gain insight into components that may influence the timing of PIPs. We drew upon the literature to formulate a theoretically-based model that illustrates processes accounting for a medication poisoning. The precursor components of age, number of unique drugs used, and number of prescription fills were identified as influencing medication poisoning. For our outcome component, we chose length of time until emergency room visit (i.e., time to failure) from the first fill of the medication. The effect of the precursor components acting individually and concurrently on the outcome component was empirically tested. In the next section, components are associated with selected hypotheses and discussion follows each hypothesis.
Conceptual background. Owing to older patients' generally poorer health status, a greater likelihood of receiving multiple prescriptions exists. Current research indicates that at least 85% of the elderly (i.e., 65 years or older) use one prescription drug, and most use more than one (Ihara, Summer & Shirey 2002). Additionally, the literature suggests that approximately 60% of prescriptions are refills (Govern 2004). However, when focusing solely on individuals age 65 and over, medication use has been found to decrease as age increases (Blackwell et al. 2008).
Older individuals are also more likely to suffer from multiple medical Conditions requiring the use of multiple prescriptions (Liu & Christensen 2002). Multiple medication use can lead to a higher risk for adverse drug events including medication poisoning (Solomon 2000). However, as patients become more knowledgeable of their medications, medication errors decrease (AMA 2005). The literature addresses duration of therapy as the length of time a patient remains on a medication (Lee et al. 2002). Familiarity with the medication, such as through continual use of the medication (i.e., increased duration of therapy), reduces the opportunity for an adverse drug event to occur (Morrow et al. 2005; Hope et al. 2004; Wong, Norton & Wittkowsky 1999). For our theoretical model, duration of therapy serves as a proxy for the PIP's familiarity with the medication. As duration of therapy increases, so does the PIP's familiarity with the medication. In the model, we assess the opportunity for an adverse event based on the time to the poisoning event following the initial fill of the offending medication by the PIP.
The following hypotheses are tested in our model:
* Hypothesis 1 (Hi): Age has a direct effect on the number of unique drugs a PIP consumes.
* Hypothesis 2 (H2): As the number of unique drugs a PIP consumes increases, the total number of prescription fills for each unique drug (i.e., duration of therapy) during the 90-day period prior to the PIP event increases.
* Hypothesis 3 (H3): The total number of differing unique drugs a PIP consumes has a direct effect on the timing of the PIP event.
* Hypothesis 4 (H4): The total number of prescription fills a PIP receives for the same unique drug (i.e., duration of therapy) has a direct effect on the timing of the PIP event.
Model Assessment. Assumptions applicable to structural equation modeling include linearity and multivariate normality (Hair et al. 1995; Bentler & Chou 1987). Transformation of the data (e.g., logarithm transformation) was used, when necessary, to support the assumption of linearity (Hair et al. 1995). Multivariate normality was assessed by calculating the normalized estimate of Mardia's multivariate kurtosis coefficient (Mardia 1970). We used the WLS method to estimate the model parameters and fit. Analyses of the data were performed using the CALLS structural equation modeling procedure (SAS Institute 2010).
Patient demographics. We found that 936 ambulatory dually eligible Medicare beneficiaries 65 years and older had a single PIP event (Table 1). Younger age (i.e., the 65-74 age group), male gender, black origin, and southern region had the highest proportion of PIPs. Over 98% of PIPs were subsequently admitted for hospitalization. The average number of unique medications consumed by a PIP within 90 days prior to the PIP event was 13 and the average number of total medication fills (i.e., original fill and refills) was 23.
PIPs by ICD-9-CM category and psychoactive/ nonpsychoactive classification are presented in Table 2. PIPs associated with psychoactives were proportionally over double that of nonpsychoactives (14.3 PIPs per 100,000 beneficiaries and 6.6 PIPs per 100,000 beneficiaries, respectively). For psychoactives, the ICD9-CM category of barbiturates and other sedatives had the highest proportion of PIPs; analgesics, antipyretics, and antirheumatics the lowest. The medication zolpidem had the highest proportion of PIPs among psychoactives. Hydrocodone was the most problematic agent among opioids and other medications comprising the analgesic, antipyretics, and antirheumatics category. For nonpsychoactives, the ICD-9-CM category of agents primarily affecting blood constituents had the highest proportion of PIPs. The medication warfarin had the highest proportion of PIPs among nonpsychoactives.
We also found that medication poisoning can neither be attributed solely to one or two ICD-9-CM categories within the psychoactive or nonpsychoactive medication classifications nor one or two medications. As shown in Table 2, we found poisonings to occur across several ICD-9-CM categories. Additionally, poisonings occurred whether the ICD-9-CM category typically contained medications having short-term use (e.g., anti-infectives) or long-term use (e.g., agents primarily affecting the cardiovascular system).
Overall, PIPs had an average of two medication fills having the same name but differing NDCs) An example of how this occurs would be a PIP receiving hydrocodone with acetaminophen or hydrochiorothiazide by two different generic drug makers. For comparative purposes, we also assessed ER visits not related to poisoning for the entire elderly dually eligible Medicare and Medicaid population for the 2003 calendar year and found the average to be one medication fill having the same medication name but a differing NDC (i.e., data not shown). Thus, PIPs were found to have an additional medication in their medication regimen having the same name but differing NDCs. However, we note that the primary reason for this difference was attributed to PIPs associated with psychoactive medications; PIPs associated with nonpsychoactive medications showed no difference (in average medication fills) from non-PIPs.
The average time to poisoning from receipt of the first medication fill for all ICD-9-CM categories was 56 days. The average for PIPs associated with psychoactives was shorter than that for PIPs with nonpsychoactives by about two days. PIPs within the nonpsychoactive categories of primarily systemic agents and anti-infectives had the lowest average time to a poisoning event (35 days and 41 days, respectively). PIPs within the nonpsychoactive medication category of agents primarily affecting the cardiovascular system had the longest average time to a poisoning event (58 days).
[FIGURE 1 OMITTED]
Table 3 presents the residual correlation matrix for the hypothesized model. Figure 1 presents the final structural model with the significant standardized regression correlations.
For hypothesis 1 ([H.sub.1]), our findings indicate very little correlation exists between the components of age and number of unique drugs filled. Age had only a slight impact on the number of unique medications filled by beneficiaries identified as PIPs. Additionally, we found the impact to be an inverse relation suggesting that as our beneficiaries age, those surviving longer tend to receive somewhat less medications than their younger counterparts.
Our findings supported hypothesis 2 ([H.sub.2[) and hypothesis 3 ([H.sub.3]). Regarding hypothesis 3 ([H.sub.3]), we found a strong association exists between the number of unique medications and time to a medication poisoning ER visit. The association is an inverse association. Thus, as the number of unique medications an elderly beneficiary receives increases (indicative of unfamiliarity with the medication), the time to a medication poisoning ER visit decreases.
Hypothesis 4 ([H.sub.4]) was also supported, Our findings suggest that a strong association exists between the total number of medication fills and time to a medication poisoning ER visit. The association is a positive association. Thus, as duration of therapy increases (indicative of increased familiarity with the medication), the time to a medication poisoning ER visit increases.
Psychoactive medications are problematic in the elderly. Potential central nervous system side effects associated with psychoactives include dizziness/vertigo, drowsiness, and/or fainting. Elderly patients are more susceptible to these side effects due to poorer health status. Additionally, with increased age, pharmacokinetics (how the body absorbs, metabolizes and eliminates a drug) and pharmacodynamics (what a drug does to the body) change. As important, the overall use of prescription drugs increases with age as older adults wrestle with the burden of chronic disease. This, in turn, leads to a greater possibility of drug-drug and drug-disease interactions being problematic for the elderly as compared to their younger counterparts. Thus, psychoactive medications such as sedatives and tranquilizers can lead to befuddlement and even drug automatism when taken by the elderly at the same dosage given to younger patients. Practitioners need to be vigilant in making dose adjustments when prescribing psychoactives to the elderly in order to compensate for these facts.
Among the psychoactives, we found barbiturates and other sedatives to be most problematic in the elderly. These findings support recommendations from previous literature. Barbiturate use in the elderly is strongly discouraged based on the Beers criteria--an explicit list of medications that the medical community suggests elderly persons should avoid (Fick et al. 2003). Regarding sedatives, between 1999 and 2006, U.S. hospital admissions due to medication poisonings from the psychoactive medications of sedatives, tranquilizers, and opioids rose from approximately 43,000 to about 71,000 (Reuters 2010). The increase was almost double the increase observed in hospitalizations for poisonings due to other medications and drugs.
The conceptual model we developed allowed us to gain insight into processes influencing poisoning events. Our findings suggest that polypharmacy (the use of multiple medications by a patient) and duration of therapy have a direct, but competing impact on the timing of a poisoning event. Past research suggests that a linear relationship exists between polypharmacy and the frequency of adverse events (Kohn, Corrigan & Donaldson 1999). Our findings add to the literature by suggesting a relationship may exist between polypharmacy and the timing of an adverse event.
Past literature also suggests that as the patient's knowledge about their medications increase, better outcomes result (Hope et al. 2004). The taking of medication has been found to be a familiar task (Morrow et al. 2005). Thus, familiarity with the medication, such as through continual use of the medication, reduces the risk of an adverse drug event (Morrow et al. 2005; Hope et al. 2004; Wong, [Norton & Wittkowsky] [Norton & Wittkowsky?] 1999). Increased reliance on the attributes of one's pills, such as pill colors and shapes, may occur if an elderly patient suffers from diminished vision due to the aging process, lacks sufficient skills to read medication labels, and/or uses many medications concurrently (Hope et al. 2004). The population in this study was the elderly dual eligible, a population comprised of low-income, frail, and disabled Medicare beneficiaries. With the ubiquity of same-strength generic products being made by differing manufacturers, the advent of differing tablet sizes, shapes, and colors abounds for any same-strength medication (produced by differing generic manufacturers). For example, a change from a blue, round tablet of warfarin 5mg from manufacturer X (received by the patient on the first prescription fill) to a yellow, oblong tablet of warfarin 5mg from manufacturer Y (received by the patient on the second prescription fill) sets the stage for a potential poisoning event to occur. Our findings add to the literature by suggesting that a relationship may exist between duration of therapy and the timing of an adverse event.
Overall, our findings suggest that it is not the confusion of old age that results in a medication poisoning, but the contributing factor of psychoactive drugs on mental status that is at work in this population. Also, polypharmacy interacting inversely with medication familiarity is a substantial contributing factor to all poisoning events, whether for psychoactive or nonpsychoactive medication use.
Lastly, our findings offer a glimpse into the challenging nature of opioid use in the elderly. Recent literature suggests that relatively little information about the comparative safety of opioid use exists (as compared to nonopioids) in treating pain (Solomon et al. 2010a, b). Current studies are only beginning to address this concern. Our findings indicate that hydrocodone, a semisynthetic opioid, was the most problematic agent within the analgesic, antipyretics, and antirheumatics category. This finding adds support to the current belief that opioids are a major drug in the growth of prescription drug poisonings (as compared to non-opioids) for pain management patients. We suggest that future efforts to monitor the trending of opioid-associated PIPS (as compared to non-opioids) should prove fruitful.
First, it should be noted that because this is a large observational study, unmeasured confounders are present. For example, medication-related interactions and comorbidities may have impacted medication poisonings. Our analysis reports findings without adjusting for confounders.
Second, conducting outcomes research on inappropriate medication use with administrative data is problematic because identified risk factors will not be randomly distributed across study groups or individuals. However, when examining large populations, the use of administrative data appears feasible as compared to the expense of primary data collection.
Third, findings based on our study population may not be fully generalizable to the population as a whole because this study examines only a particular group of beneficiaries in a particular setting (i.e., the Medicaid dual-eligible elderly population). The elderly dual eligible population is a population comprised of low-income, frail, and disabled individuals. Further research among other elderly populations will help to establish the degree to which our findings may be applied universally.
Fourth, our study is limited by the accuracy of administrative coding practices in the field and the completeness of ascertainment of medical injuries using ICD-9-CM codes. Administrative data quality in this respect compromises the quality of study results. Yet, our approach is inexpensive and quite feasible for a large population. Additionally, it is an efficient means to suggest associations for further assessment by the clinical community.
Fifth, another potential limitation is omitting important components from the model. One of the greatest weaknesses in path analysis lies in excluding key components that may influence a system. However, the use of a smaller core model will shed light on key features for development of a larger hypothesized model. Likewise, the likelihood of obtaining a good fit in the larger model increases as additional features are placed one by one into the larger model, while features of the previous smaller model may need to be replaced. For example, based on our current model, for future research, we suggest the removal of age as a feature in a larger model because of its slight impact on current model results. Additionally, because of the high volume of medication use of our identified population (i.e., patients with poisonings), we suggest adding CMS's Hierarchical Condition Category (HCC) risk score as the next key feature to the larger model through the use of more current CMS data for which HCC risk scoring is available. Furthermore, based on our current findings, we perceive submodels specific to a particular medication (e.g., hydrocodone) should be developed in order to address drug-drug interactions and drug-disease interactions specific to the particular medication associated with poisoning. Thus, we believe that for the current study, practicality was balanced with the attainment of a more comprehensive model.
Sixth, our study uses the total number of prescription fills for a unique drug as a proxy for duration of therapy. The adequacy of this as a proxy measure is potentially affected by the fact that it is sensitive to the day's supply for each prescription fill (e.g., a 30-day versus a 90-day supply for a maintenance medication). However, we perceive our methodology is an efficient means to suggest associations for further assessment by the clinical community.
The reasons behind unintentional poisoning in the ambulatory elderly have been somewhat a mystery. In this study, we initially attempt to distinguish between poisoning events associated with psychoactive medications versus nonpsychoactive medications. We found that PIPs associated with psychoactives were proportionally over double that of nonpsychoactives, with the psychoactives of barbiturates and other sedatives being the most problematic. Recent literature also suggests that relatively little information about the comparative safety of opioid use exists (as compared to non-opioids) in treating pain. Within the analgesic, antipyretics, and antirheumatics category, the drug hydrocodone, a semisynthetic opioid, was found to be the most problematic agent among agents used to treat pain.
We also operationalized differing contextual components in order to assess their impact on medication poisoning. We found that the two components of (a) use of many drugs and (b) familiarity with the medication have a direct, but competing impact on poisoning.
Based on our findings, we see an opportunity to improve prescribing behaviors of practitioners. Practitioners must remain vigilant in making dose adjustments when prescribing psychoactives to the elderly, taking into consideration that elderly patients are more susceptible to drug-induced health complications because of poorer health status, a greater potential to receive multiple medications, and differences in how the body absorbs, metabolizes and eliminates a drug.
American Medical Association (AMA). 2005. Patient safety and drug benefit: Reducing the risk of medication errors. Proceedings of the 2005 White House Conference on Aging: Mini-Conference on Health Literacy and Health Disparities. July 21. Chicago, Il.
Bentler, P.M. & Chou, C.P. 1987. Practical issues in structural modeling. Sociological Methods and Research 16 (1): 78-117.
Blackwell, S.A.; Ciborowski, G.M.; Baugh, D.K. & Montgomery, M.A. 2008. Applying the 2003 Beers update to Medicaid/Medicare enrollees. International Journal of Pharmaceutical & Healthcare Marketing 2 (4): 257-72.
Camidge, D.R.; Wood, R.J. & Bateman, D.N. 2003. The epidemiology of self-poisoning in the UK. British Journal of Clinical Pharmacology 56 (6): 613-19.
Centers for Diseases Control (CDC). 2010. Web-based injury statistics query and reporting system (WISQARS). Available at: http://www.cdc.gov/injury/wisqars/index.html.
Centers for Diseases Control (CDC). 2009. Overdose deaths involving prescription opioids among Medicaid enrollees--Washington, 2004-2007. Morbidity and Mortality Weekly Report 58 (42): 1171-75. Available at: http://www.cdc.gov/mmwr/preview/mmwrhtml/mm5842a1.htm.
Centers for Medicare & Medicaid Services (CMS) 2009. Medicaid Analytic extract (MAX) General Information. Available at: http://www.cms.gov/IdentifiableDataFiles/ 09_MedicaidAnalyticExtractMAX.asp.
Fick D.M.; Cooper J.W.; Wade W.E.; Waller J.L.; Maclean J.R. & Beers M.H. 2003. Updating the Beers Criteria for potentially inappropriate medication use in older adults. Archives of Internal Medicine 163: 2716-24.
Govern, P. 2004. Computerized prescriptions can reduce medical errors. The Reporter (Vanderbilt Medical Center) April 23. Available at: http://www.mc.vanderbilt. edu/reporter/index.html?ID=3220.
Hair, J.F.; Anderson, R.E.; Tatham, R.L. & Black, W.C. 1995. Multivariate Data Analysis with Readings. Englewood Cliffs, NJ. Prentice Hall.
Hope, C.J.; Wu, J.; Tu, W.; Young, J. & Murray, M.D. 2004. Association of medication adherence, knowledge, and skills with emergency department visits by adults 50 years or older with congestive heart failure. American Journal of Health-System Pharmacy 61 (19): 2043-49.
Ihara, E.; Summer, L. & Shirey, L. 2002. Prescription drugs: A vital component of health care. In: Challenges for the 21st Century: Chronic and Disabling Conditions. Washington, DC: Center on an Aging Society, Georgetown University. Available at: http://ihcrp.georgetown.edu/agingsociety/pdfs/rxdrugs.pdf.
Kohn, L.T.; Corrigan, J.M. & Donaldson, M.S. (Eds.) 1999. To Err is Human: Building a Safer Health System. Washington, DC: National Academy Press.
Lee, J.L.; Boro, M.S.; Knapp, K.K.; Meier, J.L. & Korman, N.E. 2002. Clinical and economic outcomes of pharmacist recommendations in a Veterans Affairs medical center. American Journal of Health-System Pharmacy 59 (21): 2070-77.
Liu, G. & Christensen, D.B. 2002. The continuing challenge of inappropriate prescribing in the elderly: An update of the evidence. Journal of the American Pharmacists Association 42 (6): 847-57.
Mardia, K.V. 1970. Measures of multivariate skewness and kurtosis with applications. Biometrika 57: 519-30.
Morrow, D.G.; Weiner, M.; Young, J.; Steinley, D.; Deer, M. & Murray, MD. 2005. Improving medication knowledge among older adults with heart failure: A patient-centered approach to instruction design. The Gerentologist 45 (4) 545-52.
National Center for Health Statistics. 2010. International Classification of Disease, Ninth Revision, Clinical Modification (ICD-9-CM). Available at: http://www.cdc.gov/nchs/icd.htm.
Paluozzi, L. & Annest, J. 2007. Unintentional poisoning deaths--United States, 1999-2004. Morbidity and Mortality Weekly Report 56 (5): 93-96.
Paulozzi, L.J.; Ballesteros, M.F. & Stevens, J.A. 2006. Recent trends in mortality from unintentional injury in the United States. Journal of Safety Research 37: 277-83.
Reuters. 2010. Prescription drug overdoses on the rise in U.S. Reuters Health April 6. Available at: http://www.reuters.com/article/id USTRE6350MR20100406
SAS Institute Inc. 2010. SAS Software version 9.1. Cary, NC: SAS Institute Inc.
Solomon, D.H.; Rassen, J.A.; Glynn, R.J.; Garneau, K.; Levin, R.; Lee, J. & Schneeweiss, S. 2010a. The comparative safety of opioids for nonmalignant pain in older adults. Archives of Internal Medicine 170: 1979-86.
Solomon, D.H.; Rassen, J.A.; Glynn, R.J.; Lee, J.; Levin, R. & Schneeweiss, S. 2010b. The comparative safety of analgesics in older adults with arthritis. Archives of Internal Medicine 170 (22): 1968-78.
Solomon, T. 2000. Pharmacology--what's new? The elderly, antidepressants, and antibiotics. Presentation at the 23rd Annual California Coalition of Nurse Practitioners Educational Conference. Anaheim, CA, March 16-19.
Wolters Kluwer Health. 2010. Master Drug Data Base. Available at: http://www.Medi-Span.com/.
Wong, W.; Norton, J.W. & Wittkowsky, A.K. 1999. Influence of warfarin regimen type on clinical and monitoring outcomes in stable patients in an anticoagulation management service. Pharmacotherapy 19 (12): 1385-91.
(1.) Medications with the same active ingredient were considered unique if differing NDCs were present. Thus, two medications identical in strength may differ in physical appearance based on color, size, shape, or packaging because they are produced by two different manufacturers.
Steven A. Blackwell, Ph.D., J.D. (a); David K. Baugh, M.A. (b); Gary M. Ciborowski, M.A. (c) & Melissa A. Montgomery, Ph.D. (d)
This article is not subject to U.S. copyright law.
This research is internally funded. The views and opinions expressed in this article are those of the authors and do not necessarily reflect the views of the CMS. The authors received input and guidance from the following individuals in the development of this article (in alphabetical order by last name): Bill Clark, Renee Meutnech, Curt Mueller, Tom Reilly, and Noelni Rudolph. The article was substantially improved by the contributions of these individuals.
(a) Social Science Research Analyst, Office of Research, Development, and Information, Centers for Medicare & Medicaid Services (CMS), Baltimore, MD.
Please address correspondence and reprint requests to Steven A. Blackwell, Office of Research, Development, and Information, Centers for Medicare & Medicaid Services, Mail Stop C3-21-28, 7500 Security Boulevard, Baltimore, MD 21244; phone: (410) 786-8652, fax: (410) 786-5610, email: firstname.lastname@example.org
(b) Senior Technical Advisor, Office of Research, Development, and Information, Centers for Medicare & Medicaid Services (CMS), Baltimore, MD.
(c) Information Technology Specialist, Office of Research, Development, and Information, Centers for Medicare & Medicaid Services (CMS), Baltimore, MD.
(d) Economist, Office of Research, Development, and Information, Centers for Medicare & Medicaid Services (CMS), Baltimore, MD.
TABLE 1 Physician Identified Poisonings (PIPs) (1) by Gender, Age Group, and Origin For Dual Enrollees Age 65 or Over in 2003 Number of PIPS Number of Dual (PIPS Per Enrollees with 100,000 Dual Number of Dual Medication Use Enrollees with Characteristic Enrollees (%) Medication Use) Gender Female 3,824,497 2,810,753 (73.5) 618 (22.0) Mule 1,588,181 1,068,286 (67.3) 318 (29.8) Total 5,412,678 3,879,039 (71.7) 936 (24.1) Age Group 65-74 2,211,392 1,520,799 (68.8) 549 (36.1) 75-84 1,981,983 1,407,729 (71.0) 299 (21.2) 85+ 1,219,3113 950,511 (78.0) 88 (9.3) Total 5,412,678 3,879,039 (71.7) 936 (24.1) Race/Ethnic Origin Black 976,033 686,439 (70.3) 223 (32.5) Hispanic 374,924 288,531 (77.0) 61 (21.1) Other 503,104 417,600 (83.0) 50 (12.0) White 3,558,617 2,486,469 (69.9) 602 (24.2) Total 5,412,678 3,879,039 (71.7) 936 (24.1) Region Midwest 1,061,474 740,320 (69.7) 162 (21.9) Northeast 1,124,531 763,356 (67.9) 138 (18.1) South 2,099,999 1,481,125 (70.5) 407 (27.5) West 1,126,674 894,238 (79.4) 229 (25.6) Total 5,412,678 3,879,039 (71.7) 936 (24.1) Average Number of Prescription Number of PIPS Fills (2) Within 90 Leading to Days Prior to ER Characteristic Hospitalization Visit for PIPS Gender Female 610 23 Mule 311 21 Total 921 23 Age Group 65-74 542 24 75-84 291 21 85+ 88 22 Total 921 23 Race/Ethnic Origin Black 220 19 Hispanic 60 18 Other 49 18 White 592 25 Total 921 23 Region Midwest 158 25 Northeast 138 23 South 400 22 West 225 21 Total 921 23 Average Number of Unique Medications (3) Filled Within 90 Days Prior to the Characteristic Poisoning Gender Female 13 Mule 12 Total 13 Age Group 65-74 13 75-84 12 85+ 12 Total 13 Race/Ethnic Origin Black 11 Hispanic 11 Other 10 White 13 Total 13 Region Midwest 14 Northeast 12 South 12 West 12 Total 13 (1) Physic ian-identified poisonings WIN are dually-eligible elderly beneficiaries that had an offending medication (as identified by the ER physician using the ICD-9-CM poisoning codes of 960-979) within 90 Clays prior to the admission. (2) Original and refill prescriptions. (3) Based on National Drug Code (NDC). TABLE 2 Physician Identified Poisonings (PIPS) (1) by ICD-9 Category for Dual Enrollees Age 65 or Older in 2003 Number of PIPS (PIPS Per 10(1,000 Beneficiaries Receiving the Problematic ICD-9-CM Category (2) Medication) Psychoactive Medications 475 (14.3) Analgesics, Antipyretics, and 86 (5.0) Antirheumatics Anticonvulsants and 97 (30.6) Anti-Parkinsonism Drugs Barbiturates and Other 62 (197.4) Sedatives (6) CNS Depressants and 6 (9.0) Anesthetics Tranqulizers, Anti-Depressants, 224 (19.2) and Anti-Psychotics (7) Non-Psychoactive Medications 461 (6.6) Anti-Infectives 6 (l.0) Hormones and Synthetic 160 (11.3) Substitutes Primarily Systemic Agents 1 (0.4) Agents Primarily Affecting 96 (14.6) Blood Constituents Agents Primarily Affecting the 159 (7.9) Cardiovascular System Agents Primarily Affecting the 3 (0.9) Gastrointestinal System Poisoning by Water, Mineral, 28 (2.2) and Uric Acid Metabolism Drugs Agents Primarily Acting on the 8 (2.0) Smooth and Skeletal Muscles and Respiratory System All Categories 936 (9.1) Average Number of Prescription Fills (3) Within 911 Days Prior to ER Visit ICD-9-CM Category (2) for PIPS Psychoactive Medications Analgesics, Antipyretics, and 27 Antirheumatics Anticonvulsants and 19 Anti-Parkinsonism Drugs Barbiturates and Other 27 Sedatives (6) CNS Depressants and 23 Anesthetics Tranqulizers, Anti-Depressants, 23 and Anti-Psychotics (7) Non-Psychoactive Medications Anti-Infectives 13 Hormones and Synthetic 22 Substitutes Primarily Systemic Agents 9 Agents Primarily Affecting 19 Blood Constituents Agents Primarily Affecting the 23 Cardiovascular System Agents Primarily Affecting the 12 Gastrointestinal System Poisoning by Water, Mineral, 27 and Uric Acid Metabolism Drugs Agents Primarily Acting on the 31 Smooth and Skeletal Muscles and Respiratory System All Categories Average Number of Medications Average With the Number of Same Name Unique but Different Medications (4) NDCs Filled Filled Within Within 90 90 Days Prior Days Prior to to ER Visit ER Visit for ICD-9-CM Category (2) for PIPS PIPS (3) Psychoactive Medications Analgesics, Antipyretics, and 15 2 Antirheumatics Anticonvulsants and 10 1 Anti-Parkinsonism Drugs Barbiturates and Other 14 1 Sedatives (6) CNS Depressants and 13 2 Anesthetics Tranqulizers, Anti-Depressants, 13 2 and Anti-Psychotics (7) Non-Psychoactive Medications Anti-Infectives 9 1 Hormones and Synthetic 12 1 Substitutes Primarily Systemic Agents 7 0 Agents Primarily Affecting 11 1 Blood Constituents Agents Primarily Affecting the 12 1 Cardiovascular System Agents Primarily Affecting the 7 0 Gastrointestinal System Poisoning by Water, Mineral, 14 1 and Uric Acid Metabolism Drugs Agents Primarily Acting on the 17 1 Smooth and Skeletal Muscles and Respiratory System All Categories Average Time to a Poisoning Event From Original Fill Within 90 Day Period Prior to ER Visit for PIPS ICD-9-CM Category (2) (in Days) Psychoactive Medications Analgesics, Antipyretics, and 55 Antirheumatics Anticonvulsants and 57 Anti-Parkinsonism Drugs Barbiturates and Other 55 Sedatives (6) CNS Depressants and 51 Anesthetics Tranqulizers, Anti-Depressants, 57 and Anti-Psychotics (7) Non-Psychoactive Medications Anti-Infectives 41 Hormones and Synthetic 57 Substitutes Primarily Systemic Agents 35 Agents Primarily Affecting 51 Blood Constituents Agents Primarily Affecting the 58 Cardiovascular System Agents Primarily Affecting the 53 Gastrointestinal System Poisoning by Water, Mineral, 48 and Uric Acid Metabolism Drugs Agents Primarily Acting on the 51 Smooth and Skeletal Muscles and Respiratory System All Categories Medication Having the Highest Number of PIPS (PIPS Per 100,000 Beneficiaries Receiving the Problematic ICD-9-CM Category (2) Medication) Psychoactive Medications Analgesics, Antipyretics, and Hydrocodone (3.2) Antirheumatics Anticonvulsants and Phenytoin (15.1) Anti-Parkinsonism Drugs Barbiturates and Other Zolpidem (5) (26.3) Sedatives (6) CNS Depressants and Carisoprodol (18.6) Anesthetics Tranqulizers, Anti-Depressants, Zolpidem (5) (16.4) and Anti-Psychotics (7) Non-Psychoactive Medications Anti-Infectives Cephalexin/Ciprofloxacin (0.8) Hormones and Synthetic Insulin (35.7) Substitutes Primarily Systemic Agents Promethazine (0.4) Agents Primarily Affecting Warfarin (41.8) Blood Constituents Agents Primarily Affecting the Amlodipine (3.8) Cardiovascular System Agents Primarily Affecting the Dicyclomine/Ranitidine/ Gastrointestinal System Lactulose (0.9) Poisoning by Water, Mineral, Furosetnide (3.4) and Uric Acid Metabolism Drugs Agents Primarily Acting on the Theophylline (4.6) Smooth and Skeletal Muscles and Respiratory System All Categories (1) Physician-identified poisonings (PIPS) are dually-eligible elderly beneficiaries that had an offending medication as identified by [lie ER physician within 90 days prior to the admission. (2) The ICD-9-CM poisoning codes of 960-979. (3) Original and refill prescriptions. (4) Based on National Drug Code (NDC). (5) Zolpidem has been classified as both a sedative and a tranquilizer. Total PIPS per 100,000 beneficiaries receiving Zolipem (as a sedative or transquilizer) are .42.6. (6) Also referred to as "Sedatives and Hypnotics" by ICD-9-CM. Only barbiturates and other sedatives were found during our analysis for this category. (7) Also referred to as "Psychotropic Agents" by ICD-9-CM. Only tranquilizers, antidepressants, and antispychotics were found during our analysis for this category. TABLE 3 Residual Correlation Matrix (1,2) Variables (3) Variables V1 V2 V3 V4 V1 0.00000 0.00192 0.00401 -0.04788 V2 0.00192 -0.00005 -0.00053 0.00442 V3 0.00401 -0.00053 0.00000 0.00495 V4 -0.04788 0.00442 0.00495 0.00000 (1) Residual matrix based on weighted least-squares estimation procedure. Residual values less than 0.10 considered acceptable. (2) Average absolute residual = 0.006377; Average off-diagonal absolute residual = 0.010620. (3) VI = Age; V2 = Number of unique medications; V3 = Total number of medication fills; V4 Time to poisoning event from the first prescription fill.
|Gale Copyright:||Copyright 2011 Gale, Cengage Learning. All rights reserved.|