View Article PDF

For medication to be therapeutically effective, it is essential that patients adhere to them. Filling a prescription is the first critical step to establish medication adherence.[[1,2]] Research on medication adherence has primarily focused on secondary non-adherence, which occurs when patients don’t take their medicines as prescribed, don’t refill their prescriptions on time or stop taking their medicines.[[3]] However, rates of primary medication non-adherence, where a patient fails to have a prescription filled for newly prescribed medicines or a suitable alternative within a specified timeframe, are far less known.[[3,4]] Failing to get newly prescribed medications filled places a burden on patients, families and the broader healthcare system by increasing mortality/morbidity, hospitalisation rates and/or emergency department (ED) visits, and it is associated with greater economic cost.[[5–7]] Previous studies have estimated that around 7–35% of patients in EDs fail to fill new prescriptions,[[6,8]]  and in paediatric EDs the reported percentage is as high as 66%.[[9]]

Non-adherence occurs due to a dysfunctional triad of patient, healthcare-system and contextual influences,[[5–7,10]] including a lack of financial and social support, the availability and/or accessibility of healthcare resources, the severity of disease and the available treatment options.[[11]] In the ED, primary medicine non-adherence is comparatively higher in patients with financial constraints.[[12–16]] Patients under financial constraints may select to have certain medicines filled over others, or may be unwilling to pay for a medicine to use in the short term.[[6,17]] In some cases, patients may also be unaware of the importance of the prescribed medicine, or may have a supply at home.[[6,17]]

Previous research found that 16% of patients discharged on ‘high importance’ medication exhibited primary non-adherence 30 days after discharge.[[7]] These prescriptions included medications such as antibiotics for the treatment of acute infections such as pneumonia, urinary tract infections and cellulitis.[[7]] Non-adherence to antimicrobial agents carries additional healthcare costs due to treatment failure, readmission to ED and resources being wasted on unused medication.[[7,18]] In the ED, adherence to antibiotics has been reported as low as 30–40%.[[18]] Analgesics are often considered to be of lesser importance compared with antibiotics, but pain is one of the most common reasons why patients visit an ED.[[19]] Although non-adherence to pain relief is generally not life-threatening, failure to receive adequate analgesia can result in significant morbidity.[[7]] Studies suggest that a significant proportion of ED prescriptions for analgesics remain unfilled[[8,19]] due to patients reporting a lack of pain, having a home supply, a fear of medication side effects or believing the analgesic is not strong enough.[[19]]

Non-adherence is also found to increase if patients are prescribed two or more medicines,[[6,16]] lack access to primary care physician,[[20]] cannot access a community pharmacy at time of discharge[[19]] or are tobacco smokers.[[11]] Patient demographic factors such as age, gender and ethnicity are also thought to influence medication adherence[[21–23]] but are inconsistently reported in the literature.[[12,16]]

In New Zealand, little is known about primary medication non-adherence in the ED. Previously we explored the relationship between non-adherence to primary medication and patient sociodemographics, smoking status, access to a regular GP and discharge time and/or day.[[24]] The aim of this study was to measure primary medication non-adherence to antibiotics, paracetamol and nonsteroidal anti-inflammatory drugs (NSAIDs) in discharge prescriptions from Counties Manukau Health Emergency Department (CMH-ED), and to determine which patient factors are likely to influence patients’ decisions to fill prescriptions for these three medication classes.

Method

This was a retrospective, observational study design of patients discharged from CMH-ED. Ethics approval was granted by the University of Auckland Human Participants Ethics Committee (Reference No. 012463) and Counties Manukau District Health Board Ethics Committee.

Participant selection

Data were collected in September from the first 1,000 patients discharged from CMH-ED between 28 April–6 May 2014, and further data were collected in January 2015 from 600 additional patients between 28 July–9 August 2014. The data were purposefully collected on these weeks to facilitate comparison between autumn and winter seasons, and the dates allowed for an assessment of adherence >90 days after the prescription order.

Data were included for patients who were residents within the Auckland Regional Public Health Service boundaries, presented to CMH-ED and were discharged with a prescription for one or more medicines. Patients were excluded if they were admitted to another ward, were transferred to another hospital, left CMH-ED without seeing a doctor or were discharged without a prescription.

Data collection

Patient data were identified through CONCERTO™ (a software programme that coordinates patient data across the whole Auckland region in a central electronic platform) by limiting the search to ‘ED speciality’ and selecting the dates required. To maintain patient confidentiality, and for auditing purposes, each patient was allocated a unique identifying code linked to their National Health Index (NHI) number. If any patients presented to the ED more than once during the study period, only their first discharge was analysed.

To facilitate data collection, a paper-based tool was developed and piloted for efficiency. Age, gender, country of birth, residential suburb, ethnicity, language spoken and details of each patient’s regular primary physician were collected if available. Presenting indication, discharge date and time, smoking status and medicines provided on discharge were obtained from electronic and paper discharge summaries. To determine whether patients had their ED prescriptions filled, information from community pharmacies was retrieved via TestSafe™, accessed via CONCERTO™. TestSafe™ records the medicines prescribed to patients via their NHI, how many medicines were filled at each dispensing, the date of dispensing and the contact details of the dispensing community pharmacy.

Data entry

Data were entered into Microsoft Excel™, and variables were coded for further analysis. Population data from Stats NZ were used to group ethnicity data into various categories. Ministry of Health guidelines assisted age categorisation, with patients under 25 years further divided into categories: under 10 years, 10–17 years (adolescent) and 18–24 years (young adults). Suburb deprivation was coded using the New Zealand Index of Deprivation (NZDep2013).[[25]] All other categories were grouped according to the information available in the patient notes. Occupation included patients under the age of five years, who were classed as ‘infants/children’, and those of school-going age (5–17 years) and/or undergoing tertiary/other education were grouped as ‘students’. Although ‘others’ were adults of employable age (18–65 years), their employment status was unclear and hence were grouped as a separate class.

Data validation

Of the total patient dataset, 20% were randomly selected using the Microsoft Excel™ randomisation function (=RANDBETWEEN) and manually cross-checked by two researchers to confirm data-entry reliability. A further 50 patients were selected by the same randomisation method and telephoned using a pre-scripted telephone checklist to enquire whether they had collected their medicine from a pharmacy following discharge. This was compared to available TestSafe™ data to ensure triangulation of the data.

Analysis

Adherence to the three medication classes of interest was analysed by the proportion of prescribed medicines that were filled. Poisson regressions were conducted with a generalised linear model with the log of the total filled per medication class per person as the offset to evaluate medication adherence. Poisson regression with one explanatory variable is given, as well as multivariable regression. Bivariable regression was added to evaluate univariable effect on adherence per medication type. All possible two-way and some relevant three-way interaction terms were evaluated to find the model of best fit. The model with the minimum Akaike Information Criterium (AIC) value was used as the best-fit multivariable model.[[26]] The evaluated independent (categorical) variables were age, ethnicity, gender, country of birth, suburb deprivation, language, occupation, regular GP, smoking, discharge date, discharge day and discharge time, medication type and total number of medications prescribed. There were no missing data in the response variables (number of filled medicines and total number prescribed per medicine). An extra level, ‘unknown’, was constructed for the missing values in the explanatory categorical variables. Priory pairwise comparisons were made to group the levels within the categorical variates, using the covariance matrix as derived from the final multivariable model. All parameters denoted by the same letter (ie, a, b, c or d) within a group are not significant from each other, and alpha is 0.05. All analyses were conducted using R (version 4.01, 64 bit).

Results

Of 1,600 patients, data were excluded for 608 because they either were discharged without a prescription (n=470; 29.4%), left without seeing a doctor (n=70; 4.4%), were non-Auckland residents (32; 2%), were admitted to a ward (n=19; 1.2%) or were transferred to another healthcare facility (n=12; 0.8%), or incomplete data were provided for them (n=5; 0.3%).

Patients were predominantly born in New Zealand (n=660; 66.5%), of Pacific Island descent (n=425; 42.8%), 24 years or younger (n=559; 56.4%) and living in the most socioeconomically deprived suburbs (NZDep2013 9 & 10) of Auckland (n=775; 78.1%). The majority were non-smokers (n=804; 81.1%) and had a regular GP (n=958; 96.6%). Almost a quarter (n=238; 23.4%) presented to the CMH-ED on a Monday, and close to half (n=481; 48.5%) were discharged between 8pm and 8am.

The majority (n=893; 90%) were discharged with a prescription for three medicines or less: of these, 29.2% (n=290) were prescribed one medication, 36.5% (n=362) were prescribed two medications and 24.3% (n=241) were prescribed three medications. The remaining 10% (n=99) were prescribed four or more medicines. Patient data were categorised by the type of medication prescribed (Table 1).

Table 1: The total number of medications prescribed and the percentage of medication types filled, as characterised across different patient variables (N=992). View Table 1.

Almost half (n=480; 48.4%) of the patient sample did not have at least one medication item on their prescription filled. Univariable analysis found filling rates significantly increased when patients were prescribed more than one medication (p≤0.01) (Table 2). Compared with paracetamol (51.3%), NSAIDs (59.9%) and other medications (61%) were significantly more likely to be filled (p=0.034 and p=0.023, respectively). Antibiotics were significantly more likely to be filled than all other medication (p<0.001). Bivariable analyses with medication type showed the strongest associations with age, smoking and number of prescribed items (Appendix Tables 1–4). Compared to patients under 10 years, patients 10–17 years were significantly less likely to have NSAIDs filled (p=0.036) and  patients 10–24 years were significantly less likely to have paracetamol prescriptions filled (p≤0.023). Smokers were significantly less adherent to paracetamol (p=0.022) and other medication (p=0.034). Paracetamol was significantly more likely to be filled if other items were also prescribed, as were other medications if three or four items were co-prescribed (p≤0.01). Patients born outside of New Zealand were more likely to fill NSAIDs (p=0.027). No significant differences were found for antibiotics. Univariable analysis has been described more fully elsewhere.[[24]]

Table 2: Univariable analysis comparing medication adherence with total number of medicines and medication types (N=992).

The best multivariable model for adherence included age, smoking and the two-way interaction medication type and number of medications prescribed (Table 3). This data confirmed that, compared with paracetamol alone, patients who were prescribed more than one item were significantly more likely to have their prescriptions filled (p<0.01) (Figure 1). Antibiotics, NSAIDs (if not the only item prescribed) and other medications were all significantly more likely to be filled compared with paracetamol (p<0.01).

Table 3: Multivariable analysis comparing medication adherence with age, smoking and number and type of medication (N=992). View Table 3.

Figure 1: Percentage (p) of medication types filled compared with number of items prescribed. Black lines indicate confidence intervals.

Discussion

In this study, 90% of patients were prescribed between one and three medications on discharge from CMH-ED, and almost half did not have at least one medication filled within 90 days. Patients were more likely to have their prescription filled when more than one medication was prescribed. Antibiotics, NSAIDs and other medications were significantly more likely to be filled compared with paracetamol alone (p<0.01). When accounting for number of medication items and type of medication, the most significant predictors for non-adherence were patients aged 10–44 years and smokers (p<0.01).

Literature has shown variable effects from age and income on whether prescriptions are filled. Some studies have revealed that older children were less likely to have their prescriptions filled,[[9]] as are those who were of low income or vulnerably housed,[[9,20]] whereas other studies have found no associations.[[6,9]] In our study, the lowest rates of prescription filling for all medication types were in children aged 10–17 years, which was significant compared to children under 10 years (p<0.001). It is important to note that, at the time of this study, the prescription co-payment charge was NZ$5 per item, a $2 increase from the previous year (2013). This co-payment was applied to all patients over the age of six years, regardless of income status,[[27]] and may have contributed to children under the age of 10 having higher prescription filling rates than older children. The co-payment increase was found to result in some patients delaying or avoiding filling their prescription and/or selecting to fill only certain medicines they deemed more important.[[27]] It has also been proposed that poorer adherence in younger patients may be due to less established or noncontinuous relationships with a primary care physician,[[11]] since in an ED setting prescribing clinicians are typically unfamiliar with the patient’s lifestyle and/or resources.[[28]] In this study, prescription filling rates of all medication types were much lower in those who did not have a regular GP, although these patient numbers were insufficient to indicate significance.

Medication-level analysis revealed an association between the medication class, number of prescribed items and likelihood of filling a prescription. Other studies have found adherence to be lower in those co-prescribed two or more medications,[[6]] or that there was no significant difference between number of medications prescribed versus the number of prescriptions filled.[[7]] Our analysis did not determine whether patients prescribed two or more medications may have been co-prescribed an antibiotic or other high-importance medication, and hence it is not possible to conclude the reasons for the higher rate of filling. It is, however, plausible that patients with more medications are sicker and/or exhibit more severe symptoms and may therefore be more inclined to fill all medications, in comparison to patients with milder and more transient presentations. Furthermore, once patients have presented to a pharmacy with a prescription, they may be more motivated to fill all items.

Paracetamol was the most commonly prescribed item (31.2% of all prescribed medications) and also the least filled prescription (51.3%). Although NSAIDs were prescribed at a similar rate to paracetamol (30.9%), NSAIDs had a significantly higher fill rate (59.9%) in comparison (p<0.01). Analgesics are reported to be some of the more common medications left unused from previous prescriptions, due largely to over-prescribing and/or the patient experiencing adverse effects.[[29]] Hence it is not uncommon for patients to have a supply of analgesics at home, which they may opt to take rather than filling another prescription. Or patients may make rational decisions to not take analgesics when they no longer seek pain relief.[[30,31]] It has been proposed that pain relief is a desired health endpoint that health professionals should not attempt to modify.[[30]] However, assessing the patient’s analgesic requirements and questioning them regarding their home supply may help to better optimise analgesic prescribing in EDs.

Unlike analgesics and anti-inflammatories, the failure to take antibiotics can lead to more potentially serious consequences. In this study, antibiotics accounted for 11.9% of prescription items and exhibited a significantly higher fill rate compared with other medications (80.4%, p<0.001). One study reported similar fill rates but found no significant differences between antibiotics and other classes of medication,[[31]] whereas other studies have found that oral anti-infective agents exhibit the lowest rates of primary medication non-adherence.[[4–6]] These discrepancies are likely due to the varying methods of dispensing antibiotics in ED studies: for example, some were dispensed in the form of a fully-paid prescription, and in others as a starter pack with instructions for a follow-on prescription.[[18]] Although in our study there were no patient factors that significantly affected antibiotic filling, there was some indication to suggest that age, ethnicity and access to a regular GP may be associated with primary medication non-adherence.

Several strategies to improve antibiotic filling rates in EDs have yielded variable results. Patients who were dispensed antibiotics directly from an ED tended to have better adherence than patients issued with a prescription, even if their prescription had been fully paid for.[[18]] Previous studies have reported that, once patients fill their prescriptions and get into a routine, they are more likely to continue taking their antibiotics;[[30]] for those who do not fill their prescriptions, commonly reported reasons were cost, lack of transportation and being busy.[[30]] It has also been suggested that patients who are discharged during pharmacy opening hours may be more likely to fill their prescriptions,[[18]] but discharge day and time was not seen to influence filling in our study.

Limitations

This study was conducted at a single ED site (CMH-ED), and given the small cohort and patient demographic at CMH-ED, it may not be possible to generalise the results to other EDs across New Zealand. Sampling of the study population was obtained from consecutive patients over a period of four weeks, which may have contributed to selection bias. As primary medication non-adherence is influenced by a multitude of factors, including ED-system factors, access to a dispensing pharmacy, patient social and psychological influences and financial support, many of these factors could not be determined using a retrospective method. Moreover, it could not be determined whether patients had an existing supply of medicine at home, experienced symptom resolution and no longer required analgesics, or were provided with a ‘back-pocket prescription’ (ie, ‘just in case’ prescribing) that was no longer required, resulting in unfilled prescription medications. Additionally, patients could not be assessed for possible medication adverse effects over time, and readmission rates due to non-adherence were not assessed.

Since medication adherence was inferred from a prescription being filled, true nonadherence rates are likely to be underestimated. TestSafe™ displayed only those prescriptions that had been entered into the system, but it did not indicate whether the medication had been collected or taken by the patient. Telephoning patients several months after ED discharge to confirm medical records could have resulted in recall bias. With two of 15 patients having reported conflicting dispensing records, it suggests that the system may have a fairly substantial error rate. These differences, however, could also be based on inaccurate recall or biased by social desirability. Moreover, admission records for CMH-ED patients were not always accurate or complete. In busy EDs, these are often written under time constraints and personal information is mainly obtained through patient self-reporting.

Conclusions

In this study, age, smoking and number of prescribed medications were predictors of non-adherence to antibiotics and analgesics. Since this study, changes have been made to the New Zealand prescription co-payment structure and free paediatric prescriptions have been extended to 13-year-olds. There are also a growing number of discount pharmacies offering free prescriptions for all patients. With the cost of medication affecting the ability of some patients to fill their prescriptions, free prescriptions may ease the financial burden for people who are unable to afford co-payments and improve treatment opportunities, and further research is warranted to determine whether these changes have any notable effects on the rates of non-adherence in the ED. More work is required to identify patients likely to be non-adherent and enlist available resources to reduce barriers to adherence. Hospital discharge may be the best time at which to communicate medication treatments and emphasise medication adherence.

Appendix

Appendix Table 1: Bivariable analysis of paracetamol (N=992). View Appendix Table 1.

Appendix Table 2: Bivariable analysis of nonsteroidal anti-inflammatory drugs (N=992). View Appendix Table 2.

Appendix Table 3: Bivariable analysis of antibiotics (N=992). View Appendix Table 3.

Appendix Table 4: Bivariable analysis of other medication (N=992). View Appendix Table 4.

Summary

Abstract

AIM: To measure primary medication non-adherence to antibiotics, paracetamol and nonsteroidal anti-inflammatory drugs (NSAIDs) in patients discharged from Counties Manukau Health Emergency Department (CMH-ED). METHOD: A retrospective observational study based on 1,600 discharged patients’ data collected between 28 April–6 May and 28 July–9 August 2014. Data were included for patients who were residents within the Auckland Regional Public Health Service boundaries, presented to CMH-ED and were discharged with a prescription. RESULTS: Of 992 patients, 48.5% did not have at least one medication on their discharge prescription filled. Patients were mostly born in New Zealand (66.5%), of Pacific Island descent (42.8%), living in the most socioeconomically deprived areas (78.1%) and under 10 years of age (32.6%). Filling rates significantly increased with >1 prescribed item (p≤0.01). NSAIDs were significantly more likely to be filled compared with paracetamol (59.9% vs 51.3%, p=0.034); antibiotics were significantly more likely to be filled than all other medicines (80.4%, p<0.001). The most significant predictors for non-adherence when accounting for number and types of medications were patients 10–44 years (p<0.05) and smokers (p<0.01). CONCLUSIONS: Age, smoking and number of prescribed medications were predictors of non-adherence to medication type. Further research is warranted to assess whether changes to prescription co-payments affect the rate of nonadherence.

Aim

Method

Results

Conclusion

Author Information

Nataly Martini: Senior Lecturer, School of Pharmacy, Faculty of Medical and Health Sciences, Auckland, New Zealand. Bert van der Werf: Senior Research Fellow, Department of Epidemiology and Biostatistics, School of Population Health, Faculty of Medical and Health Sciences, Auckland, New Zealand. Deborah Bassett-Clarke: Teacher-Practitioner, School of Pharmacy, Faculty of Medical and Health Sciences, Auckland, New Zealand.

Acknowledgements

We thank Tarik Al-Diery, Shannon Carey, Shannon Chow, Salja Kumar, Joanna Liang, Aaron Naidoo and Hamish Wu for assistance with data collection, and Professor Gregory Luke Larkin for guidance. We also gratefully acknowledge the support offered by the School of Pharmacy, The University of Auckland, for travel and printing costs.

Correspondence

Nataly Martini, Private bag 92019, Auckland, New Zealand, +6499232150

Correspondence Email

n.martini@auckland.ac.nz

Competing Interests

Nil.

1. Wamala S, Merlo J, Bostrom G, Hogstedt C, Agren G. Socioeconomic disadvantage and primary non-adherence with medication in Sweden. Int J Qual Health Care. 2007 Jun 1;19(3):134–40.

2. Matsui D, Joubert GIE, Dykxhoorn S, Rieder MJ. Compliance With Prescription Filling in the Pediatric Emergency Department. Arch Pediatr Adolesc Med. 2000 Feb 1;154(2):195–8.

3. Lee S-Q, Raamkumar AS, Li J, Cao Y, Witedwittayanusat K, Chen L, et al. Reasons for Primary Medication Nonadherence: A Systematic Review and Metric Analysis. J Manag Care Spec Pharm. 2018 Jul 30;24(8):778–94.

4. Fischer MA, Stedman MR, Lii J, Vogeli C, Shrank WH, Brookhart MA, et al. Primary Medication Non-Adherence: Analysis of 195,930 Electronic Prescriptions. J Gen Intern Med N Y. 2010 Apr;25(4):284–90.

5. Hensley C, Heaton PC, Kahn RS, Luder HR, Frede SM, Beck AF. Poverty, Transportation Access, and Medication Nonadherence. Pediatrics. 2018 Apr;141(4):e20173402.

6. Hohl CM, Abu-Laban RB, Brubacher JR, Zed PJ, Sobolev B, Tsai G, et al. Adherence to emergency department discharge prescriptions. CJEM Can J Emerg Med. 2009 Mar;11(2):131–8.

7. Fallis BA, Dhalla IA, Klemensberg J, Bell CM. Primary medication non-adherence after discharge from a general internal medicine service. PloS One. 2013;8(5):e61735.

8. Farris B, Shakowski C, Mueller SW, Phong S, Kiser TH, Jacknin G. Patient nonadherence to filling discharge medication prescriptions from the emergency department: Barriers and clinical implications. Am J Health Syst Pharm. 2018 Mar 1;75(5):316–20.

9. Rosman SL, Dorfman D, Suglia SF, Humphrey C, Silverstein M. Predictors of prescription filling after visits to the pediatric emergency department. Pediatr Emerg Care. 2012 Jan;28(1):22–5.

10. Aggarwal B, Mosca L. Lifestyle and Psychosocial Risk Factors Predict Non-adherence to Medication. Ann Behav Med Publ Soc Behav Med. 2010 Oct;40(2):228–33.

11. Feehan M, Morrison MA, Tak C, Morisky DE, DeAngelis MM, Munger MA. Factors predicting self-reported medication low adherence in a large sample of adults in the US general population: a cross-sectional study. BMJ Open. 2017 Jun;7(6):e014435.

12. Davis DP, Jandrisevits MD, Iles S, Weber TR, Gallo LC. Demographic, Socioeconomic, and Psychological Factors Related to Medication Non-adherence Among Emergency Department Patients. J Emerg Med. 2012 Nov 1;43(5):773–85.

13. Thomas EJ, Burstin HR, O’Neil AC, Orav EJ, Brennan TA. Patient Noncompliance With Medical Advice After the Emergency Department Visit. Ann Emerg Med. 1996 Jan 1;27(1):49–55.

14. Reed M, Brand R, Newhouse JP, Selby JV, Hsu J. Coping with Prescription Drug Cost Sharing: Knowledge, Adherence, and Financial Burden. Health Serv Res. 2008;43(2):785–97.

15. Klein D, Turvey C, Wallace R. Elders Who Delay Medication Because of Cost: Health Insurance, Demographic, Health, and Financial Correlates. The Gerontologist. 2004 Dec 1;44(6):779–87.

16. Heaton PC, Tundia NL, Luder HR. U.S. emergency departments visits resulting from poor medication adherence: 2005–07. J Am Pharm Assoc. 2013 Sep 1;53(5):513–9.

17. Jones I, Britten N. Why do some patients not cash their prescriptions? Br J Gen Pract. 1998 Jan;48(426):903–5.

18. Ho J, Taylor DM, Cabalag MS, Ugoni A, Yeoh M. Factors that impact on emergency department patient compliance with antibiotic regimens. Emerg Med J. 2010 Nov 1;27(11):815–20.

19. Mcintosh SE, Leffler S. Pain management after discharge from the ED. Am J Emerg Med. 2004 Mar 1;22(2):98–100.

20. Hunter CE, Palepu A, Farrell S, Gogosis E, O’Brien K, Hwang SW. Barriers to Prescription Medication Adherence Among Homeless and Vulnerably Housed Adults in Three Canadian Cities: J Prim Care Community Health [Internet]. 2014 Nov 26 [cited 2020 Aug 23]; Available from: https://journals.sagepub.com/doi/10.1177/2150131914560610?url_ver=Z39.88-2003&rfr_id=ori%3Arid%3Acrossref.org&rfr_dat=cr_pub++0pubmed

21. Fischer MA, Choudhry NK, Brill G, Avorn J, Schneeweiss S, Hutchins D, et al. Trouble Getting Started: Predictors of Primary Medication Nonadherence. Am J Med. 2011 Nov 1;124(11):1081.e9-1081.e22.

22. Bazargan M, Barbre AR, Hamm V. Failure to have Prescriptions Filled among Black Elderly. J Aging Health. 1993 May 1;5(2):264–82.

23. Kennedy J, Erb C. Prescription noncompliance due to cost among adults with disabilities in the United States. Am J Public Health Wash. 2002 Jul;92(7):1120–4.

24. Martini ND, Werf B van der, Bassett-Clarke D. Primary medication non-adherence at Counties Manukau Health Emergency Department (CMH-ED), New Zealand: an observational study. BMJ Open. 2020 Jul 1;10(7):e035775.

25. NZDep2013 Index of Deprivation [Internet]. Ministry of Health NZ. [cited 2019 Oct 25]. Available from: https://www.health.govt.nz/publication/nzdep2013-index-deprivation

26. Burnham KP, Anderson DR, editors. Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach. New York, NY: Springer New York; 2002

27. Norris P, Tordoff J, McIntosh B, Laxman K, Chang SY, Te Karu L. Impact of prescription charges on people living in poverty: A qualitative study. Res Soc Adm Pharm. 2016 Nov 1;12(6):893–902.

28. McCarthy ML, Ding R, Roderer NK, Steinwachs DM, Ortmann MJ, Pham JC, et al. Does Providing Prescription Information or Services Improve Medication Adherence Among Patients Discharged From the Emergency Department? A Randomized Controlled Trial. Ann Emerg Med. 2013 Sep;62(3):212-223.e1.

29. Maeng DD, Tom LA, Wright EA. Patient characteristics and healthcare utilization patterns associated with unused medications among medicare patients. Res Soc Adm Pharm. 2017 Nov 1;13(6):1090–4.

30. Suffoletto B, Yealy DM. The Trouble With Medication Adherence After Emergency Care. Ann Emerg Med. 2013 Sep 1;62(3):235–6.

31. Sammons NW, Yin H. Compliance of medications prescribed from a pediatric emergency department. Pediatr Emerg Care. 2015 Jun;31(6):399–402.

For the PDF of this article,
contact nzmj@nzma.org.nz

View Article PDF

For medication to be therapeutically effective, it is essential that patients adhere to them. Filling a prescription is the first critical step to establish medication adherence.[[1,2]] Research on medication adherence has primarily focused on secondary non-adherence, which occurs when patients don’t take their medicines as prescribed, don’t refill their prescriptions on time or stop taking their medicines.[[3]] However, rates of primary medication non-adherence, where a patient fails to have a prescription filled for newly prescribed medicines or a suitable alternative within a specified timeframe, are far less known.[[3,4]] Failing to get newly prescribed medications filled places a burden on patients, families and the broader healthcare system by increasing mortality/morbidity, hospitalisation rates and/or emergency department (ED) visits, and it is associated with greater economic cost.[[5–7]] Previous studies have estimated that around 7–35% of patients in EDs fail to fill new prescriptions,[[6,8]]  and in paediatric EDs the reported percentage is as high as 66%.[[9]]

Non-adherence occurs due to a dysfunctional triad of patient, healthcare-system and contextual influences,[[5–7,10]] including a lack of financial and social support, the availability and/or accessibility of healthcare resources, the severity of disease and the available treatment options.[[11]] In the ED, primary medicine non-adherence is comparatively higher in patients with financial constraints.[[12–16]] Patients under financial constraints may select to have certain medicines filled over others, or may be unwilling to pay for a medicine to use in the short term.[[6,17]] In some cases, patients may also be unaware of the importance of the prescribed medicine, or may have a supply at home.[[6,17]]

Previous research found that 16% of patients discharged on ‘high importance’ medication exhibited primary non-adherence 30 days after discharge.[[7]] These prescriptions included medications such as antibiotics for the treatment of acute infections such as pneumonia, urinary tract infections and cellulitis.[[7]] Non-adherence to antimicrobial agents carries additional healthcare costs due to treatment failure, readmission to ED and resources being wasted on unused medication.[[7,18]] In the ED, adherence to antibiotics has been reported as low as 30–40%.[[18]] Analgesics are often considered to be of lesser importance compared with antibiotics, but pain is one of the most common reasons why patients visit an ED.[[19]] Although non-adherence to pain relief is generally not life-threatening, failure to receive adequate analgesia can result in significant morbidity.[[7]] Studies suggest that a significant proportion of ED prescriptions for analgesics remain unfilled[[8,19]] due to patients reporting a lack of pain, having a home supply, a fear of medication side effects or believing the analgesic is not strong enough.[[19]]

Non-adherence is also found to increase if patients are prescribed two or more medicines,[[6,16]] lack access to primary care physician,[[20]] cannot access a community pharmacy at time of discharge[[19]] or are tobacco smokers.[[11]] Patient demographic factors such as age, gender and ethnicity are also thought to influence medication adherence[[21–23]] but are inconsistently reported in the literature.[[12,16]]

In New Zealand, little is known about primary medication non-adherence in the ED. Previously we explored the relationship between non-adherence to primary medication and patient sociodemographics, smoking status, access to a regular GP and discharge time and/or day.[[24]] The aim of this study was to measure primary medication non-adherence to antibiotics, paracetamol and nonsteroidal anti-inflammatory drugs (NSAIDs) in discharge prescriptions from Counties Manukau Health Emergency Department (CMH-ED), and to determine which patient factors are likely to influence patients’ decisions to fill prescriptions for these three medication classes.

Method

This was a retrospective, observational study design of patients discharged from CMH-ED. Ethics approval was granted by the University of Auckland Human Participants Ethics Committee (Reference No. 012463) and Counties Manukau District Health Board Ethics Committee.

Participant selection

Data were collected in September from the first 1,000 patients discharged from CMH-ED between 28 April–6 May 2014, and further data were collected in January 2015 from 600 additional patients between 28 July–9 August 2014. The data were purposefully collected on these weeks to facilitate comparison between autumn and winter seasons, and the dates allowed for an assessment of adherence >90 days after the prescription order.

Data were included for patients who were residents within the Auckland Regional Public Health Service boundaries, presented to CMH-ED and were discharged with a prescription for one or more medicines. Patients were excluded if they were admitted to another ward, were transferred to another hospital, left CMH-ED without seeing a doctor or were discharged without a prescription.

Data collection

Patient data were identified through CONCERTO™ (a software programme that coordinates patient data across the whole Auckland region in a central electronic platform) by limiting the search to ‘ED speciality’ and selecting the dates required. To maintain patient confidentiality, and for auditing purposes, each patient was allocated a unique identifying code linked to their National Health Index (NHI) number. If any patients presented to the ED more than once during the study period, only their first discharge was analysed.

To facilitate data collection, a paper-based tool was developed and piloted for efficiency. Age, gender, country of birth, residential suburb, ethnicity, language spoken and details of each patient’s regular primary physician were collected if available. Presenting indication, discharge date and time, smoking status and medicines provided on discharge were obtained from electronic and paper discharge summaries. To determine whether patients had their ED prescriptions filled, information from community pharmacies was retrieved via TestSafe™, accessed via CONCERTO™. TestSafe™ records the medicines prescribed to patients via their NHI, how many medicines were filled at each dispensing, the date of dispensing and the contact details of the dispensing community pharmacy.

Data entry

Data were entered into Microsoft Excel™, and variables were coded for further analysis. Population data from Stats NZ were used to group ethnicity data into various categories. Ministry of Health guidelines assisted age categorisation, with patients under 25 years further divided into categories: under 10 years, 10–17 years (adolescent) and 18–24 years (young adults). Suburb deprivation was coded using the New Zealand Index of Deprivation (NZDep2013).[[25]] All other categories were grouped according to the information available in the patient notes. Occupation included patients under the age of five years, who were classed as ‘infants/children’, and those of school-going age (5–17 years) and/or undergoing tertiary/other education were grouped as ‘students’. Although ‘others’ were adults of employable age (18–65 years), their employment status was unclear and hence were grouped as a separate class.

Data validation

Of the total patient dataset, 20% were randomly selected using the Microsoft Excel™ randomisation function (=RANDBETWEEN) and manually cross-checked by two researchers to confirm data-entry reliability. A further 50 patients were selected by the same randomisation method and telephoned using a pre-scripted telephone checklist to enquire whether they had collected their medicine from a pharmacy following discharge. This was compared to available TestSafe™ data to ensure triangulation of the data.

Analysis

Adherence to the three medication classes of interest was analysed by the proportion of prescribed medicines that were filled. Poisson regressions were conducted with a generalised linear model with the log of the total filled per medication class per person as the offset to evaluate medication adherence. Poisson regression with one explanatory variable is given, as well as multivariable regression. Bivariable regression was added to evaluate univariable effect on adherence per medication type. All possible two-way and some relevant three-way interaction terms were evaluated to find the model of best fit. The model with the minimum Akaike Information Criterium (AIC) value was used as the best-fit multivariable model.[[26]] The evaluated independent (categorical) variables were age, ethnicity, gender, country of birth, suburb deprivation, language, occupation, regular GP, smoking, discharge date, discharge day and discharge time, medication type and total number of medications prescribed. There were no missing data in the response variables (number of filled medicines and total number prescribed per medicine). An extra level, ‘unknown’, was constructed for the missing values in the explanatory categorical variables. Priory pairwise comparisons were made to group the levels within the categorical variates, using the covariance matrix as derived from the final multivariable model. All parameters denoted by the same letter (ie, a, b, c or d) within a group are not significant from each other, and alpha is 0.05. All analyses were conducted using R (version 4.01, 64 bit).

Results

Of 1,600 patients, data were excluded for 608 because they either were discharged without a prescription (n=470; 29.4%), left without seeing a doctor (n=70; 4.4%), were non-Auckland residents (32; 2%), were admitted to a ward (n=19; 1.2%) or were transferred to another healthcare facility (n=12; 0.8%), or incomplete data were provided for them (n=5; 0.3%).

Patients were predominantly born in New Zealand (n=660; 66.5%), of Pacific Island descent (n=425; 42.8%), 24 years or younger (n=559; 56.4%) and living in the most socioeconomically deprived suburbs (NZDep2013 9 & 10) of Auckland (n=775; 78.1%). The majority were non-smokers (n=804; 81.1%) and had a regular GP (n=958; 96.6%). Almost a quarter (n=238; 23.4%) presented to the CMH-ED on a Monday, and close to half (n=481; 48.5%) were discharged between 8pm and 8am.

The majority (n=893; 90%) were discharged with a prescription for three medicines or less: of these, 29.2% (n=290) were prescribed one medication, 36.5% (n=362) were prescribed two medications and 24.3% (n=241) were prescribed three medications. The remaining 10% (n=99) were prescribed four or more medicines. Patient data were categorised by the type of medication prescribed (Table 1).

Table 1: The total number of medications prescribed and the percentage of medication types filled, as characterised across different patient variables (N=992). View Table 1.

Almost half (n=480; 48.4%) of the patient sample did not have at least one medication item on their prescription filled. Univariable analysis found filling rates significantly increased when patients were prescribed more than one medication (p≤0.01) (Table 2). Compared with paracetamol (51.3%), NSAIDs (59.9%) and other medications (61%) were significantly more likely to be filled (p=0.034 and p=0.023, respectively). Antibiotics were significantly more likely to be filled than all other medication (p<0.001). Bivariable analyses with medication type showed the strongest associations with age, smoking and number of prescribed items (Appendix Tables 1–4). Compared to patients under 10 years, patients 10–17 years were significantly less likely to have NSAIDs filled (p=0.036) and  patients 10–24 years were significantly less likely to have paracetamol prescriptions filled (p≤0.023). Smokers were significantly less adherent to paracetamol (p=0.022) and other medication (p=0.034). Paracetamol was significantly more likely to be filled if other items were also prescribed, as were other medications if three or four items were co-prescribed (p≤0.01). Patients born outside of New Zealand were more likely to fill NSAIDs (p=0.027). No significant differences were found for antibiotics. Univariable analysis has been described more fully elsewhere.[[24]]

Table 2: Univariable analysis comparing medication adherence with total number of medicines and medication types (N=992).

The best multivariable model for adherence included age, smoking and the two-way interaction medication type and number of medications prescribed (Table 3). This data confirmed that, compared with paracetamol alone, patients who were prescribed more than one item were significantly more likely to have their prescriptions filled (p<0.01) (Figure 1). Antibiotics, NSAIDs (if not the only item prescribed) and other medications were all significantly more likely to be filled compared with paracetamol (p<0.01).

Table 3: Multivariable analysis comparing medication adherence with age, smoking and number and type of medication (N=992). View Table 3.

Figure 1: Percentage (p) of medication types filled compared with number of items prescribed. Black lines indicate confidence intervals.

Discussion

In this study, 90% of patients were prescribed between one and three medications on discharge from CMH-ED, and almost half did not have at least one medication filled within 90 days. Patients were more likely to have their prescription filled when more than one medication was prescribed. Antibiotics, NSAIDs and other medications were significantly more likely to be filled compared with paracetamol alone (p<0.01). When accounting for number of medication items and type of medication, the most significant predictors for non-adherence were patients aged 10–44 years and smokers (p<0.01).

Literature has shown variable effects from age and income on whether prescriptions are filled. Some studies have revealed that older children were less likely to have their prescriptions filled,[[9]] as are those who were of low income or vulnerably housed,[[9,20]] whereas other studies have found no associations.[[6,9]] In our study, the lowest rates of prescription filling for all medication types were in children aged 10–17 years, which was significant compared to children under 10 years (p<0.001). It is important to note that, at the time of this study, the prescription co-payment charge was NZ$5 per item, a $2 increase from the previous year (2013). This co-payment was applied to all patients over the age of six years, regardless of income status,[[27]] and may have contributed to children under the age of 10 having higher prescription filling rates than older children. The co-payment increase was found to result in some patients delaying or avoiding filling their prescription and/or selecting to fill only certain medicines they deemed more important.[[27]] It has also been proposed that poorer adherence in younger patients may be due to less established or noncontinuous relationships with a primary care physician,[[11]] since in an ED setting prescribing clinicians are typically unfamiliar with the patient’s lifestyle and/or resources.[[28]] In this study, prescription filling rates of all medication types were much lower in those who did not have a regular GP, although these patient numbers were insufficient to indicate significance.

Medication-level analysis revealed an association between the medication class, number of prescribed items and likelihood of filling a prescription. Other studies have found adherence to be lower in those co-prescribed two or more medications,[[6]] or that there was no significant difference between number of medications prescribed versus the number of prescriptions filled.[[7]] Our analysis did not determine whether patients prescribed two or more medications may have been co-prescribed an antibiotic or other high-importance medication, and hence it is not possible to conclude the reasons for the higher rate of filling. It is, however, plausible that patients with more medications are sicker and/or exhibit more severe symptoms and may therefore be more inclined to fill all medications, in comparison to patients with milder and more transient presentations. Furthermore, once patients have presented to a pharmacy with a prescription, they may be more motivated to fill all items.

Paracetamol was the most commonly prescribed item (31.2% of all prescribed medications) and also the least filled prescription (51.3%). Although NSAIDs were prescribed at a similar rate to paracetamol (30.9%), NSAIDs had a significantly higher fill rate (59.9%) in comparison (p<0.01). Analgesics are reported to be some of the more common medications left unused from previous prescriptions, due largely to over-prescribing and/or the patient experiencing adverse effects.[[29]] Hence it is not uncommon for patients to have a supply of analgesics at home, which they may opt to take rather than filling another prescription. Or patients may make rational decisions to not take analgesics when they no longer seek pain relief.[[30,31]] It has been proposed that pain relief is a desired health endpoint that health professionals should not attempt to modify.[[30]] However, assessing the patient’s analgesic requirements and questioning them regarding their home supply may help to better optimise analgesic prescribing in EDs.

Unlike analgesics and anti-inflammatories, the failure to take antibiotics can lead to more potentially serious consequences. In this study, antibiotics accounted for 11.9% of prescription items and exhibited a significantly higher fill rate compared with other medications (80.4%, p<0.001). One study reported similar fill rates but found no significant differences between antibiotics and other classes of medication,[[31]] whereas other studies have found that oral anti-infective agents exhibit the lowest rates of primary medication non-adherence.[[4–6]] These discrepancies are likely due to the varying methods of dispensing antibiotics in ED studies: for example, some were dispensed in the form of a fully-paid prescription, and in others as a starter pack with instructions for a follow-on prescription.[[18]] Although in our study there were no patient factors that significantly affected antibiotic filling, there was some indication to suggest that age, ethnicity and access to a regular GP may be associated with primary medication non-adherence.

Several strategies to improve antibiotic filling rates in EDs have yielded variable results. Patients who were dispensed antibiotics directly from an ED tended to have better adherence than patients issued with a prescription, even if their prescription had been fully paid for.[[18]] Previous studies have reported that, once patients fill their prescriptions and get into a routine, they are more likely to continue taking their antibiotics;[[30]] for those who do not fill their prescriptions, commonly reported reasons were cost, lack of transportation and being busy.[[30]] It has also been suggested that patients who are discharged during pharmacy opening hours may be more likely to fill their prescriptions,[[18]] but discharge day and time was not seen to influence filling in our study.

Limitations

This study was conducted at a single ED site (CMH-ED), and given the small cohort and patient demographic at CMH-ED, it may not be possible to generalise the results to other EDs across New Zealand. Sampling of the study population was obtained from consecutive patients over a period of four weeks, which may have contributed to selection bias. As primary medication non-adherence is influenced by a multitude of factors, including ED-system factors, access to a dispensing pharmacy, patient social and psychological influences and financial support, many of these factors could not be determined using a retrospective method. Moreover, it could not be determined whether patients had an existing supply of medicine at home, experienced symptom resolution and no longer required analgesics, or were provided with a ‘back-pocket prescription’ (ie, ‘just in case’ prescribing) that was no longer required, resulting in unfilled prescription medications. Additionally, patients could not be assessed for possible medication adverse effects over time, and readmission rates due to non-adherence were not assessed.

Since medication adherence was inferred from a prescription being filled, true nonadherence rates are likely to be underestimated. TestSafe™ displayed only those prescriptions that had been entered into the system, but it did not indicate whether the medication had been collected or taken by the patient. Telephoning patients several months after ED discharge to confirm medical records could have resulted in recall bias. With two of 15 patients having reported conflicting dispensing records, it suggests that the system may have a fairly substantial error rate. These differences, however, could also be based on inaccurate recall or biased by social desirability. Moreover, admission records for CMH-ED patients were not always accurate or complete. In busy EDs, these are often written under time constraints and personal information is mainly obtained through patient self-reporting.

Conclusions

In this study, age, smoking and number of prescribed medications were predictors of non-adherence to antibiotics and analgesics. Since this study, changes have been made to the New Zealand prescription co-payment structure and free paediatric prescriptions have been extended to 13-year-olds. There are also a growing number of discount pharmacies offering free prescriptions for all patients. With the cost of medication affecting the ability of some patients to fill their prescriptions, free prescriptions may ease the financial burden for people who are unable to afford co-payments and improve treatment opportunities, and further research is warranted to determine whether these changes have any notable effects on the rates of non-adherence in the ED. More work is required to identify patients likely to be non-adherent and enlist available resources to reduce barriers to adherence. Hospital discharge may be the best time at which to communicate medication treatments and emphasise medication adherence.

Appendix

Appendix Table 1: Bivariable analysis of paracetamol (N=992). View Appendix Table 1.

Appendix Table 2: Bivariable analysis of nonsteroidal anti-inflammatory drugs (N=992). View Appendix Table 2.

Appendix Table 3: Bivariable analysis of antibiotics (N=992). View Appendix Table 3.

Appendix Table 4: Bivariable analysis of other medication (N=992). View Appendix Table 4.

Summary

Abstract

AIM: To measure primary medication non-adherence to antibiotics, paracetamol and nonsteroidal anti-inflammatory drugs (NSAIDs) in patients discharged from Counties Manukau Health Emergency Department (CMH-ED). METHOD: A retrospective observational study based on 1,600 discharged patients’ data collected between 28 April–6 May and 28 July–9 August 2014. Data were included for patients who were residents within the Auckland Regional Public Health Service boundaries, presented to CMH-ED and were discharged with a prescription. RESULTS: Of 992 patients, 48.5% did not have at least one medication on their discharge prescription filled. Patients were mostly born in New Zealand (66.5%), of Pacific Island descent (42.8%), living in the most socioeconomically deprived areas (78.1%) and under 10 years of age (32.6%). Filling rates significantly increased with >1 prescribed item (p≤0.01). NSAIDs were significantly more likely to be filled compared with paracetamol (59.9% vs 51.3%, p=0.034); antibiotics were significantly more likely to be filled than all other medicines (80.4%, p<0.001). The most significant predictors for non-adherence when accounting for number and types of medications were patients 10–44 years (p<0.05) and smokers (p<0.01). CONCLUSIONS: Age, smoking and number of prescribed medications were predictors of non-adherence to medication type. Further research is warranted to assess whether changes to prescription co-payments affect the rate of nonadherence.

Aim

Method

Results

Conclusion

Author Information

Nataly Martini: Senior Lecturer, School of Pharmacy, Faculty of Medical and Health Sciences, Auckland, New Zealand. Bert van der Werf: Senior Research Fellow, Department of Epidemiology and Biostatistics, School of Population Health, Faculty of Medical and Health Sciences, Auckland, New Zealand. Deborah Bassett-Clarke: Teacher-Practitioner, School of Pharmacy, Faculty of Medical and Health Sciences, Auckland, New Zealand.

Acknowledgements

We thank Tarik Al-Diery, Shannon Carey, Shannon Chow, Salja Kumar, Joanna Liang, Aaron Naidoo and Hamish Wu for assistance with data collection, and Professor Gregory Luke Larkin for guidance. We also gratefully acknowledge the support offered by the School of Pharmacy, The University of Auckland, for travel and printing costs.

Correspondence

Nataly Martini, Private bag 92019, Auckland, New Zealand, +6499232150

Correspondence Email

n.martini@auckland.ac.nz

Competing Interests

Nil.

1. Wamala S, Merlo J, Bostrom G, Hogstedt C, Agren G. Socioeconomic disadvantage and primary non-adherence with medication in Sweden. Int J Qual Health Care. 2007 Jun 1;19(3):134–40.

2. Matsui D, Joubert GIE, Dykxhoorn S, Rieder MJ. Compliance With Prescription Filling in the Pediatric Emergency Department. Arch Pediatr Adolesc Med. 2000 Feb 1;154(2):195–8.

3. Lee S-Q, Raamkumar AS, Li J, Cao Y, Witedwittayanusat K, Chen L, et al. Reasons for Primary Medication Nonadherence: A Systematic Review and Metric Analysis. J Manag Care Spec Pharm. 2018 Jul 30;24(8):778–94.

4. Fischer MA, Stedman MR, Lii J, Vogeli C, Shrank WH, Brookhart MA, et al. Primary Medication Non-Adherence: Analysis of 195,930 Electronic Prescriptions. J Gen Intern Med N Y. 2010 Apr;25(4):284–90.

5. Hensley C, Heaton PC, Kahn RS, Luder HR, Frede SM, Beck AF. Poverty, Transportation Access, and Medication Nonadherence. Pediatrics. 2018 Apr;141(4):e20173402.

6. Hohl CM, Abu-Laban RB, Brubacher JR, Zed PJ, Sobolev B, Tsai G, et al. Adherence to emergency department discharge prescriptions. CJEM Can J Emerg Med. 2009 Mar;11(2):131–8.

7. Fallis BA, Dhalla IA, Klemensberg J, Bell CM. Primary medication non-adherence after discharge from a general internal medicine service. PloS One. 2013;8(5):e61735.

8. Farris B, Shakowski C, Mueller SW, Phong S, Kiser TH, Jacknin G. Patient nonadherence to filling discharge medication prescriptions from the emergency department: Barriers and clinical implications. Am J Health Syst Pharm. 2018 Mar 1;75(5):316–20.

9. Rosman SL, Dorfman D, Suglia SF, Humphrey C, Silverstein M. Predictors of prescription filling after visits to the pediatric emergency department. Pediatr Emerg Care. 2012 Jan;28(1):22–5.

10. Aggarwal B, Mosca L. Lifestyle and Psychosocial Risk Factors Predict Non-adherence to Medication. Ann Behav Med Publ Soc Behav Med. 2010 Oct;40(2):228–33.

11. Feehan M, Morrison MA, Tak C, Morisky DE, DeAngelis MM, Munger MA. Factors predicting self-reported medication low adherence in a large sample of adults in the US general population: a cross-sectional study. BMJ Open. 2017 Jun;7(6):e014435.

12. Davis DP, Jandrisevits MD, Iles S, Weber TR, Gallo LC. Demographic, Socioeconomic, and Psychological Factors Related to Medication Non-adherence Among Emergency Department Patients. J Emerg Med. 2012 Nov 1;43(5):773–85.

13. Thomas EJ, Burstin HR, O’Neil AC, Orav EJ, Brennan TA. Patient Noncompliance With Medical Advice After the Emergency Department Visit. Ann Emerg Med. 1996 Jan 1;27(1):49–55.

14. Reed M, Brand R, Newhouse JP, Selby JV, Hsu J. Coping with Prescription Drug Cost Sharing: Knowledge, Adherence, and Financial Burden. Health Serv Res. 2008;43(2):785–97.

15. Klein D, Turvey C, Wallace R. Elders Who Delay Medication Because of Cost: Health Insurance, Demographic, Health, and Financial Correlates. The Gerontologist. 2004 Dec 1;44(6):779–87.

16. Heaton PC, Tundia NL, Luder HR. U.S. emergency departments visits resulting from poor medication adherence: 2005–07. J Am Pharm Assoc. 2013 Sep 1;53(5):513–9.

17. Jones I, Britten N. Why do some patients not cash their prescriptions? Br J Gen Pract. 1998 Jan;48(426):903–5.

18. Ho J, Taylor DM, Cabalag MS, Ugoni A, Yeoh M. Factors that impact on emergency department patient compliance with antibiotic regimens. Emerg Med J. 2010 Nov 1;27(11):815–20.

19. Mcintosh SE, Leffler S. Pain management after discharge from the ED. Am J Emerg Med. 2004 Mar 1;22(2):98–100.

20. Hunter CE, Palepu A, Farrell S, Gogosis E, O’Brien K, Hwang SW. Barriers to Prescription Medication Adherence Among Homeless and Vulnerably Housed Adults in Three Canadian Cities: J Prim Care Community Health [Internet]. 2014 Nov 26 [cited 2020 Aug 23]; Available from: https://journals.sagepub.com/doi/10.1177/2150131914560610?url_ver=Z39.88-2003&rfr_id=ori%3Arid%3Acrossref.org&rfr_dat=cr_pub++0pubmed

21. Fischer MA, Choudhry NK, Brill G, Avorn J, Schneeweiss S, Hutchins D, et al. Trouble Getting Started: Predictors of Primary Medication Nonadherence. Am J Med. 2011 Nov 1;124(11):1081.e9-1081.e22.

22. Bazargan M, Barbre AR, Hamm V. Failure to have Prescriptions Filled among Black Elderly. J Aging Health. 1993 May 1;5(2):264–82.

23. Kennedy J, Erb C. Prescription noncompliance due to cost among adults with disabilities in the United States. Am J Public Health Wash. 2002 Jul;92(7):1120–4.

24. Martini ND, Werf B van der, Bassett-Clarke D. Primary medication non-adherence at Counties Manukau Health Emergency Department (CMH-ED), New Zealand: an observational study. BMJ Open. 2020 Jul 1;10(7):e035775.

25. NZDep2013 Index of Deprivation [Internet]. Ministry of Health NZ. [cited 2019 Oct 25]. Available from: https://www.health.govt.nz/publication/nzdep2013-index-deprivation

26. Burnham KP, Anderson DR, editors. Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach. New York, NY: Springer New York; 2002

27. Norris P, Tordoff J, McIntosh B, Laxman K, Chang SY, Te Karu L. Impact of prescription charges on people living in poverty: A qualitative study. Res Soc Adm Pharm. 2016 Nov 1;12(6):893–902.

28. McCarthy ML, Ding R, Roderer NK, Steinwachs DM, Ortmann MJ, Pham JC, et al. Does Providing Prescription Information or Services Improve Medication Adherence Among Patients Discharged From the Emergency Department? A Randomized Controlled Trial. Ann Emerg Med. 2013 Sep;62(3):212-223.e1.

29. Maeng DD, Tom LA, Wright EA. Patient characteristics and healthcare utilization patterns associated with unused medications among medicare patients. Res Soc Adm Pharm. 2017 Nov 1;13(6):1090–4.

30. Suffoletto B, Yealy DM. The Trouble With Medication Adherence After Emergency Care. Ann Emerg Med. 2013 Sep 1;62(3):235–6.

31. Sammons NW, Yin H. Compliance of medications prescribed from a pediatric emergency department. Pediatr Emerg Care. 2015 Jun;31(6):399–402.

For the PDF of this article,
contact nzmj@nzma.org.nz

View Article PDF

For medication to be therapeutically effective, it is essential that patients adhere to them. Filling a prescription is the first critical step to establish medication adherence.[[1,2]] Research on medication adherence has primarily focused on secondary non-adherence, which occurs when patients don’t take their medicines as prescribed, don’t refill their prescriptions on time or stop taking their medicines.[[3]] However, rates of primary medication non-adherence, where a patient fails to have a prescription filled for newly prescribed medicines or a suitable alternative within a specified timeframe, are far less known.[[3,4]] Failing to get newly prescribed medications filled places a burden on patients, families and the broader healthcare system by increasing mortality/morbidity, hospitalisation rates and/or emergency department (ED) visits, and it is associated with greater economic cost.[[5–7]] Previous studies have estimated that around 7–35% of patients in EDs fail to fill new prescriptions,[[6,8]]  and in paediatric EDs the reported percentage is as high as 66%.[[9]]

Non-adherence occurs due to a dysfunctional triad of patient, healthcare-system and contextual influences,[[5–7,10]] including a lack of financial and social support, the availability and/or accessibility of healthcare resources, the severity of disease and the available treatment options.[[11]] In the ED, primary medicine non-adherence is comparatively higher in patients with financial constraints.[[12–16]] Patients under financial constraints may select to have certain medicines filled over others, or may be unwilling to pay for a medicine to use in the short term.[[6,17]] In some cases, patients may also be unaware of the importance of the prescribed medicine, or may have a supply at home.[[6,17]]

Previous research found that 16% of patients discharged on ‘high importance’ medication exhibited primary non-adherence 30 days after discharge.[[7]] These prescriptions included medications such as antibiotics for the treatment of acute infections such as pneumonia, urinary tract infections and cellulitis.[[7]] Non-adherence to antimicrobial agents carries additional healthcare costs due to treatment failure, readmission to ED and resources being wasted on unused medication.[[7,18]] In the ED, adherence to antibiotics has been reported as low as 30–40%.[[18]] Analgesics are often considered to be of lesser importance compared with antibiotics, but pain is one of the most common reasons why patients visit an ED.[[19]] Although non-adherence to pain relief is generally not life-threatening, failure to receive adequate analgesia can result in significant morbidity.[[7]] Studies suggest that a significant proportion of ED prescriptions for analgesics remain unfilled[[8,19]] due to patients reporting a lack of pain, having a home supply, a fear of medication side effects or believing the analgesic is not strong enough.[[19]]

Non-adherence is also found to increase if patients are prescribed two or more medicines,[[6,16]] lack access to primary care physician,[[20]] cannot access a community pharmacy at time of discharge[[19]] or are tobacco smokers.[[11]] Patient demographic factors such as age, gender and ethnicity are also thought to influence medication adherence[[21–23]] but are inconsistently reported in the literature.[[12,16]]

In New Zealand, little is known about primary medication non-adherence in the ED. Previously we explored the relationship between non-adherence to primary medication and patient sociodemographics, smoking status, access to a regular GP and discharge time and/or day.[[24]] The aim of this study was to measure primary medication non-adherence to antibiotics, paracetamol and nonsteroidal anti-inflammatory drugs (NSAIDs) in discharge prescriptions from Counties Manukau Health Emergency Department (CMH-ED), and to determine which patient factors are likely to influence patients’ decisions to fill prescriptions for these three medication classes.

Method

This was a retrospective, observational study design of patients discharged from CMH-ED. Ethics approval was granted by the University of Auckland Human Participants Ethics Committee (Reference No. 012463) and Counties Manukau District Health Board Ethics Committee.

Participant selection

Data were collected in September from the first 1,000 patients discharged from CMH-ED between 28 April–6 May 2014, and further data were collected in January 2015 from 600 additional patients between 28 July–9 August 2014. The data were purposefully collected on these weeks to facilitate comparison between autumn and winter seasons, and the dates allowed for an assessment of adherence >90 days after the prescription order.

Data were included for patients who were residents within the Auckland Regional Public Health Service boundaries, presented to CMH-ED and were discharged with a prescription for one or more medicines. Patients were excluded if they were admitted to another ward, were transferred to another hospital, left CMH-ED without seeing a doctor or were discharged without a prescription.

Data collection

Patient data were identified through CONCERTO™ (a software programme that coordinates patient data across the whole Auckland region in a central electronic platform) by limiting the search to ‘ED speciality’ and selecting the dates required. To maintain patient confidentiality, and for auditing purposes, each patient was allocated a unique identifying code linked to their National Health Index (NHI) number. If any patients presented to the ED more than once during the study period, only their first discharge was analysed.

To facilitate data collection, a paper-based tool was developed and piloted for efficiency. Age, gender, country of birth, residential suburb, ethnicity, language spoken and details of each patient’s regular primary physician were collected if available. Presenting indication, discharge date and time, smoking status and medicines provided on discharge were obtained from electronic and paper discharge summaries. To determine whether patients had their ED prescriptions filled, information from community pharmacies was retrieved via TestSafe™, accessed via CONCERTO™. TestSafe™ records the medicines prescribed to patients via their NHI, how many medicines were filled at each dispensing, the date of dispensing and the contact details of the dispensing community pharmacy.

Data entry

Data were entered into Microsoft Excel™, and variables were coded for further analysis. Population data from Stats NZ were used to group ethnicity data into various categories. Ministry of Health guidelines assisted age categorisation, with patients under 25 years further divided into categories: under 10 years, 10–17 years (adolescent) and 18–24 years (young adults). Suburb deprivation was coded using the New Zealand Index of Deprivation (NZDep2013).[[25]] All other categories were grouped according to the information available in the patient notes. Occupation included patients under the age of five years, who were classed as ‘infants/children’, and those of school-going age (5–17 years) and/or undergoing tertiary/other education were grouped as ‘students’. Although ‘others’ were adults of employable age (18–65 years), their employment status was unclear and hence were grouped as a separate class.

Data validation

Of the total patient dataset, 20% were randomly selected using the Microsoft Excel™ randomisation function (=RANDBETWEEN) and manually cross-checked by two researchers to confirm data-entry reliability. A further 50 patients were selected by the same randomisation method and telephoned using a pre-scripted telephone checklist to enquire whether they had collected their medicine from a pharmacy following discharge. This was compared to available TestSafe™ data to ensure triangulation of the data.

Analysis

Adherence to the three medication classes of interest was analysed by the proportion of prescribed medicines that were filled. Poisson regressions were conducted with a generalised linear model with the log of the total filled per medication class per person as the offset to evaluate medication adherence. Poisson regression with one explanatory variable is given, as well as multivariable regression. Bivariable regression was added to evaluate univariable effect on adherence per medication type. All possible two-way and some relevant three-way interaction terms were evaluated to find the model of best fit. The model with the minimum Akaike Information Criterium (AIC) value was used as the best-fit multivariable model.[[26]] The evaluated independent (categorical) variables were age, ethnicity, gender, country of birth, suburb deprivation, language, occupation, regular GP, smoking, discharge date, discharge day and discharge time, medication type and total number of medications prescribed. There were no missing data in the response variables (number of filled medicines and total number prescribed per medicine). An extra level, ‘unknown’, was constructed for the missing values in the explanatory categorical variables. Priory pairwise comparisons were made to group the levels within the categorical variates, using the covariance matrix as derived from the final multivariable model. All parameters denoted by the same letter (ie, a, b, c or d) within a group are not significant from each other, and alpha is 0.05. All analyses were conducted using R (version 4.01, 64 bit).

Results

Of 1,600 patients, data were excluded for 608 because they either were discharged without a prescription (n=470; 29.4%), left without seeing a doctor (n=70; 4.4%), were non-Auckland residents (32; 2%), were admitted to a ward (n=19; 1.2%) or were transferred to another healthcare facility (n=12; 0.8%), or incomplete data were provided for them (n=5; 0.3%).

Patients were predominantly born in New Zealand (n=660; 66.5%), of Pacific Island descent (n=425; 42.8%), 24 years or younger (n=559; 56.4%) and living in the most socioeconomically deprived suburbs (NZDep2013 9 & 10) of Auckland (n=775; 78.1%). The majority were non-smokers (n=804; 81.1%) and had a regular GP (n=958; 96.6%). Almost a quarter (n=238; 23.4%) presented to the CMH-ED on a Monday, and close to half (n=481; 48.5%) were discharged between 8pm and 8am.

The majority (n=893; 90%) were discharged with a prescription for three medicines or less: of these, 29.2% (n=290) were prescribed one medication, 36.5% (n=362) were prescribed two medications and 24.3% (n=241) were prescribed three medications. The remaining 10% (n=99) were prescribed four or more medicines. Patient data were categorised by the type of medication prescribed (Table 1).

Table 1: The total number of medications prescribed and the percentage of medication types filled, as characterised across different patient variables (N=992). View Table 1.

Almost half (n=480; 48.4%) of the patient sample did not have at least one medication item on their prescription filled. Univariable analysis found filling rates significantly increased when patients were prescribed more than one medication (p≤0.01) (Table 2). Compared with paracetamol (51.3%), NSAIDs (59.9%) and other medications (61%) were significantly more likely to be filled (p=0.034 and p=0.023, respectively). Antibiotics were significantly more likely to be filled than all other medication (p<0.001). Bivariable analyses with medication type showed the strongest associations with age, smoking and number of prescribed items (Appendix Tables 1–4). Compared to patients under 10 years, patients 10–17 years were significantly less likely to have NSAIDs filled (p=0.036) and  patients 10–24 years were significantly less likely to have paracetamol prescriptions filled (p≤0.023). Smokers were significantly less adherent to paracetamol (p=0.022) and other medication (p=0.034). Paracetamol was significantly more likely to be filled if other items were also prescribed, as were other medications if three or four items were co-prescribed (p≤0.01). Patients born outside of New Zealand were more likely to fill NSAIDs (p=0.027). No significant differences were found for antibiotics. Univariable analysis has been described more fully elsewhere.[[24]]

Table 2: Univariable analysis comparing medication adherence with total number of medicines and medication types (N=992).

The best multivariable model for adherence included age, smoking and the two-way interaction medication type and number of medications prescribed (Table 3). This data confirmed that, compared with paracetamol alone, patients who were prescribed more than one item were significantly more likely to have their prescriptions filled (p<0.01) (Figure 1). Antibiotics, NSAIDs (if not the only item prescribed) and other medications were all significantly more likely to be filled compared with paracetamol (p<0.01).

Table 3: Multivariable analysis comparing medication adherence with age, smoking and number and type of medication (N=992). View Table 3.

Figure 1: Percentage (p) of medication types filled compared with number of items prescribed. Black lines indicate confidence intervals.

Discussion

In this study, 90% of patients were prescribed between one and three medications on discharge from CMH-ED, and almost half did not have at least one medication filled within 90 days. Patients were more likely to have their prescription filled when more than one medication was prescribed. Antibiotics, NSAIDs and other medications were significantly more likely to be filled compared with paracetamol alone (p<0.01). When accounting for number of medication items and type of medication, the most significant predictors for non-adherence were patients aged 10–44 years and smokers (p<0.01).

Literature has shown variable effects from age and income on whether prescriptions are filled. Some studies have revealed that older children were less likely to have their prescriptions filled,[[9]] as are those who were of low income or vulnerably housed,[[9,20]] whereas other studies have found no associations.[[6,9]] In our study, the lowest rates of prescription filling for all medication types were in children aged 10–17 years, which was significant compared to children under 10 years (p<0.001). It is important to note that, at the time of this study, the prescription co-payment charge was NZ$5 per item, a $2 increase from the previous year (2013). This co-payment was applied to all patients over the age of six years, regardless of income status,[[27]] and may have contributed to children under the age of 10 having higher prescription filling rates than older children. The co-payment increase was found to result in some patients delaying or avoiding filling their prescription and/or selecting to fill only certain medicines they deemed more important.[[27]] It has also been proposed that poorer adherence in younger patients may be due to less established or noncontinuous relationships with a primary care physician,[[11]] since in an ED setting prescribing clinicians are typically unfamiliar with the patient’s lifestyle and/or resources.[[28]] In this study, prescription filling rates of all medication types were much lower in those who did not have a regular GP, although these patient numbers were insufficient to indicate significance.

Medication-level analysis revealed an association between the medication class, number of prescribed items and likelihood of filling a prescription. Other studies have found adherence to be lower in those co-prescribed two or more medications,[[6]] or that there was no significant difference between number of medications prescribed versus the number of prescriptions filled.[[7]] Our analysis did not determine whether patients prescribed two or more medications may have been co-prescribed an antibiotic or other high-importance medication, and hence it is not possible to conclude the reasons for the higher rate of filling. It is, however, plausible that patients with more medications are sicker and/or exhibit more severe symptoms and may therefore be more inclined to fill all medications, in comparison to patients with milder and more transient presentations. Furthermore, once patients have presented to a pharmacy with a prescription, they may be more motivated to fill all items.

Paracetamol was the most commonly prescribed item (31.2% of all prescribed medications) and also the least filled prescription (51.3%). Although NSAIDs were prescribed at a similar rate to paracetamol (30.9%), NSAIDs had a significantly higher fill rate (59.9%) in comparison (p<0.01). Analgesics are reported to be some of the more common medications left unused from previous prescriptions, due largely to over-prescribing and/or the patient experiencing adverse effects.[[29]] Hence it is not uncommon for patients to have a supply of analgesics at home, which they may opt to take rather than filling another prescription. Or patients may make rational decisions to not take analgesics when they no longer seek pain relief.[[30,31]] It has been proposed that pain relief is a desired health endpoint that health professionals should not attempt to modify.[[30]] However, assessing the patient’s analgesic requirements and questioning them regarding their home supply may help to better optimise analgesic prescribing in EDs.

Unlike analgesics and anti-inflammatories, the failure to take antibiotics can lead to more potentially serious consequences. In this study, antibiotics accounted for 11.9% of prescription items and exhibited a significantly higher fill rate compared with other medications (80.4%, p<0.001). One study reported similar fill rates but found no significant differences between antibiotics and other classes of medication,[[31]] whereas other studies have found that oral anti-infective agents exhibit the lowest rates of primary medication non-adherence.[[4–6]] These discrepancies are likely due to the varying methods of dispensing antibiotics in ED studies: for example, some were dispensed in the form of a fully-paid prescription, and in others as a starter pack with instructions for a follow-on prescription.[[18]] Although in our study there were no patient factors that significantly affected antibiotic filling, there was some indication to suggest that age, ethnicity and access to a regular GP may be associated with primary medication non-adherence.

Several strategies to improve antibiotic filling rates in EDs have yielded variable results. Patients who were dispensed antibiotics directly from an ED tended to have better adherence than patients issued with a prescription, even if their prescription had been fully paid for.[[18]] Previous studies have reported that, once patients fill their prescriptions and get into a routine, they are more likely to continue taking their antibiotics;[[30]] for those who do not fill their prescriptions, commonly reported reasons were cost, lack of transportation and being busy.[[30]] It has also been suggested that patients who are discharged during pharmacy opening hours may be more likely to fill their prescriptions,[[18]] but discharge day and time was not seen to influence filling in our study.

Limitations

This study was conducted at a single ED site (CMH-ED), and given the small cohort and patient demographic at CMH-ED, it may not be possible to generalise the results to other EDs across New Zealand. Sampling of the study population was obtained from consecutive patients over a period of four weeks, which may have contributed to selection bias. As primary medication non-adherence is influenced by a multitude of factors, including ED-system factors, access to a dispensing pharmacy, patient social and psychological influences and financial support, many of these factors could not be determined using a retrospective method. Moreover, it could not be determined whether patients had an existing supply of medicine at home, experienced symptom resolution and no longer required analgesics, or were provided with a ‘back-pocket prescription’ (ie, ‘just in case’ prescribing) that was no longer required, resulting in unfilled prescription medications. Additionally, patients could not be assessed for possible medication adverse effects over time, and readmission rates due to non-adherence were not assessed.

Since medication adherence was inferred from a prescription being filled, true nonadherence rates are likely to be underestimated. TestSafe™ displayed only those prescriptions that had been entered into the system, but it did not indicate whether the medication had been collected or taken by the patient. Telephoning patients several months after ED discharge to confirm medical records could have resulted in recall bias. With two of 15 patients having reported conflicting dispensing records, it suggests that the system may have a fairly substantial error rate. These differences, however, could also be based on inaccurate recall or biased by social desirability. Moreover, admission records for CMH-ED patients were not always accurate or complete. In busy EDs, these are often written under time constraints and personal information is mainly obtained through patient self-reporting.

Conclusions

In this study, age, smoking and number of prescribed medications were predictors of non-adherence to antibiotics and analgesics. Since this study, changes have been made to the New Zealand prescription co-payment structure and free paediatric prescriptions have been extended to 13-year-olds. There are also a growing number of discount pharmacies offering free prescriptions for all patients. With the cost of medication affecting the ability of some patients to fill their prescriptions, free prescriptions may ease the financial burden for people who are unable to afford co-payments and improve treatment opportunities, and further research is warranted to determine whether these changes have any notable effects on the rates of non-adherence in the ED. More work is required to identify patients likely to be non-adherent and enlist available resources to reduce barriers to adherence. Hospital discharge may be the best time at which to communicate medication treatments and emphasise medication adherence.

Appendix

Appendix Table 1: Bivariable analysis of paracetamol (N=992). View Appendix Table 1.

Appendix Table 2: Bivariable analysis of nonsteroidal anti-inflammatory drugs (N=992). View Appendix Table 2.

Appendix Table 3: Bivariable analysis of antibiotics (N=992). View Appendix Table 3.

Appendix Table 4: Bivariable analysis of other medication (N=992). View Appendix Table 4.

Summary

Abstract

AIM: To measure primary medication non-adherence to antibiotics, paracetamol and nonsteroidal anti-inflammatory drugs (NSAIDs) in patients discharged from Counties Manukau Health Emergency Department (CMH-ED). METHOD: A retrospective observational study based on 1,600 discharged patients’ data collected between 28 April–6 May and 28 July–9 August 2014. Data were included for patients who were residents within the Auckland Regional Public Health Service boundaries, presented to CMH-ED and were discharged with a prescription. RESULTS: Of 992 patients, 48.5% did not have at least one medication on their discharge prescription filled. Patients were mostly born in New Zealand (66.5%), of Pacific Island descent (42.8%), living in the most socioeconomically deprived areas (78.1%) and under 10 years of age (32.6%). Filling rates significantly increased with >1 prescribed item (p≤0.01). NSAIDs were significantly more likely to be filled compared with paracetamol (59.9% vs 51.3%, p=0.034); antibiotics were significantly more likely to be filled than all other medicines (80.4%, p<0.001). The most significant predictors for non-adherence when accounting for number and types of medications were patients 10–44 years (p<0.05) and smokers (p<0.01). CONCLUSIONS: Age, smoking and number of prescribed medications were predictors of non-adherence to medication type. Further research is warranted to assess whether changes to prescription co-payments affect the rate of nonadherence.

Aim

Method

Results

Conclusion

Author Information

Nataly Martini: Senior Lecturer, School of Pharmacy, Faculty of Medical and Health Sciences, Auckland, New Zealand. Bert van der Werf: Senior Research Fellow, Department of Epidemiology and Biostatistics, School of Population Health, Faculty of Medical and Health Sciences, Auckland, New Zealand. Deborah Bassett-Clarke: Teacher-Practitioner, School of Pharmacy, Faculty of Medical and Health Sciences, Auckland, New Zealand.

Acknowledgements

We thank Tarik Al-Diery, Shannon Carey, Shannon Chow, Salja Kumar, Joanna Liang, Aaron Naidoo and Hamish Wu for assistance with data collection, and Professor Gregory Luke Larkin for guidance. We also gratefully acknowledge the support offered by the School of Pharmacy, The University of Auckland, for travel and printing costs.

Correspondence

Nataly Martini, Private bag 92019, Auckland, New Zealand, +6499232150

Correspondence Email

n.martini@auckland.ac.nz

Competing Interests

Nil.

1. Wamala S, Merlo J, Bostrom G, Hogstedt C, Agren G. Socioeconomic disadvantage and primary non-adherence with medication in Sweden. Int J Qual Health Care. 2007 Jun 1;19(3):134–40.

2. Matsui D, Joubert GIE, Dykxhoorn S, Rieder MJ. Compliance With Prescription Filling in the Pediatric Emergency Department. Arch Pediatr Adolesc Med. 2000 Feb 1;154(2):195–8.

3. Lee S-Q, Raamkumar AS, Li J, Cao Y, Witedwittayanusat K, Chen L, et al. Reasons for Primary Medication Nonadherence: A Systematic Review and Metric Analysis. J Manag Care Spec Pharm. 2018 Jul 30;24(8):778–94.

4. Fischer MA, Stedman MR, Lii J, Vogeli C, Shrank WH, Brookhart MA, et al. Primary Medication Non-Adherence: Analysis of 195,930 Electronic Prescriptions. J Gen Intern Med N Y. 2010 Apr;25(4):284–90.

5. Hensley C, Heaton PC, Kahn RS, Luder HR, Frede SM, Beck AF. Poverty, Transportation Access, and Medication Nonadherence. Pediatrics. 2018 Apr;141(4):e20173402.

6. Hohl CM, Abu-Laban RB, Brubacher JR, Zed PJ, Sobolev B, Tsai G, et al. Adherence to emergency department discharge prescriptions. CJEM Can J Emerg Med. 2009 Mar;11(2):131–8.

7. Fallis BA, Dhalla IA, Klemensberg J, Bell CM. Primary medication non-adherence after discharge from a general internal medicine service. PloS One. 2013;8(5):e61735.

8. Farris B, Shakowski C, Mueller SW, Phong S, Kiser TH, Jacknin G. Patient nonadherence to filling discharge medication prescriptions from the emergency department: Barriers and clinical implications. Am J Health Syst Pharm. 2018 Mar 1;75(5):316–20.

9. Rosman SL, Dorfman D, Suglia SF, Humphrey C, Silverstein M. Predictors of prescription filling after visits to the pediatric emergency department. Pediatr Emerg Care. 2012 Jan;28(1):22–5.

10. Aggarwal B, Mosca L. Lifestyle and Psychosocial Risk Factors Predict Non-adherence to Medication. Ann Behav Med Publ Soc Behav Med. 2010 Oct;40(2):228–33.

11. Feehan M, Morrison MA, Tak C, Morisky DE, DeAngelis MM, Munger MA. Factors predicting self-reported medication low adherence in a large sample of adults in the US general population: a cross-sectional study. BMJ Open. 2017 Jun;7(6):e014435.

12. Davis DP, Jandrisevits MD, Iles S, Weber TR, Gallo LC. Demographic, Socioeconomic, and Psychological Factors Related to Medication Non-adherence Among Emergency Department Patients. J Emerg Med. 2012 Nov 1;43(5):773–85.

13. Thomas EJ, Burstin HR, O’Neil AC, Orav EJ, Brennan TA. Patient Noncompliance With Medical Advice After the Emergency Department Visit. Ann Emerg Med. 1996 Jan 1;27(1):49–55.

14. Reed M, Brand R, Newhouse JP, Selby JV, Hsu J. Coping with Prescription Drug Cost Sharing: Knowledge, Adherence, and Financial Burden. Health Serv Res. 2008;43(2):785–97.

15. Klein D, Turvey C, Wallace R. Elders Who Delay Medication Because of Cost: Health Insurance, Demographic, Health, and Financial Correlates. The Gerontologist. 2004 Dec 1;44(6):779–87.

16. Heaton PC, Tundia NL, Luder HR. U.S. emergency departments visits resulting from poor medication adherence: 2005–07. J Am Pharm Assoc. 2013 Sep 1;53(5):513–9.

17. Jones I, Britten N. Why do some patients not cash their prescriptions? Br J Gen Pract. 1998 Jan;48(426):903–5.

18. Ho J, Taylor DM, Cabalag MS, Ugoni A, Yeoh M. Factors that impact on emergency department patient compliance with antibiotic regimens. Emerg Med J. 2010 Nov 1;27(11):815–20.

19. Mcintosh SE, Leffler S. Pain management after discharge from the ED. Am J Emerg Med. 2004 Mar 1;22(2):98–100.

20. Hunter CE, Palepu A, Farrell S, Gogosis E, O’Brien K, Hwang SW. Barriers to Prescription Medication Adherence Among Homeless and Vulnerably Housed Adults in Three Canadian Cities: J Prim Care Community Health [Internet]. 2014 Nov 26 [cited 2020 Aug 23]; Available from: https://journals.sagepub.com/doi/10.1177/2150131914560610?url_ver=Z39.88-2003&rfr_id=ori%3Arid%3Acrossref.org&rfr_dat=cr_pub++0pubmed

21. Fischer MA, Choudhry NK, Brill G, Avorn J, Schneeweiss S, Hutchins D, et al. Trouble Getting Started: Predictors of Primary Medication Nonadherence. Am J Med. 2011 Nov 1;124(11):1081.e9-1081.e22.

22. Bazargan M, Barbre AR, Hamm V. Failure to have Prescriptions Filled among Black Elderly. J Aging Health. 1993 May 1;5(2):264–82.

23. Kennedy J, Erb C. Prescription noncompliance due to cost among adults with disabilities in the United States. Am J Public Health Wash. 2002 Jul;92(7):1120–4.

24. Martini ND, Werf B van der, Bassett-Clarke D. Primary medication non-adherence at Counties Manukau Health Emergency Department (CMH-ED), New Zealand: an observational study. BMJ Open. 2020 Jul 1;10(7):e035775.

25. NZDep2013 Index of Deprivation [Internet]. Ministry of Health NZ. [cited 2019 Oct 25]. Available from: https://www.health.govt.nz/publication/nzdep2013-index-deprivation

26. Burnham KP, Anderson DR, editors. Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach. New York, NY: Springer New York; 2002

27. Norris P, Tordoff J, McIntosh B, Laxman K, Chang SY, Te Karu L. Impact of prescription charges on people living in poverty: A qualitative study. Res Soc Adm Pharm. 2016 Nov 1;12(6):893–902.

28. McCarthy ML, Ding R, Roderer NK, Steinwachs DM, Ortmann MJ, Pham JC, et al. Does Providing Prescription Information or Services Improve Medication Adherence Among Patients Discharged From the Emergency Department? A Randomized Controlled Trial. Ann Emerg Med. 2013 Sep;62(3):212-223.e1.

29. Maeng DD, Tom LA, Wright EA. Patient characteristics and healthcare utilization patterns associated with unused medications among medicare patients. Res Soc Adm Pharm. 2017 Nov 1;13(6):1090–4.

30. Suffoletto B, Yealy DM. The Trouble With Medication Adherence After Emergency Care. Ann Emerg Med. 2013 Sep 1;62(3):235–6.

31. Sammons NW, Yin H. Compliance of medications prescribed from a pediatric emergency department. Pediatr Emerg Care. 2015 Jun;31(6):399–402.

Contact diana@nzma.org.nz
for the PDF of this article

Subscriber Content

The full contents of this pages only available to subscribers.

LOGINSUBSCRIBE