Medication safety - the introduction and evaluation of interventions- Bryony Dean Franklin Professor of Medication Safety, UCL School of Pharmacy Director, Centre for Medication Safety and Service Quality, Imperial College Healthcare NHS Trust Chair, Imperial Centre for Patient Safety and Service Quality Associate Editor, BMJ Quality and Safety

Most common healthcare intervention… Use of medication is the most common intervention healthcare - nearly every patient admitted to hospital will be prescribed medication, and a third are prescribed anti-infectives at any one time. Two thirds of patients registered with a GP will receive at least one prescription in a year, and about 12% of these are for anti-infectives.

But… International systematic reviews: median prescribing error rate: 7.0% of inpatient medication orders 1 Median medication administration error rate: 8.0% doses, excluding wrong time errors 2 Median 3.7% of unplanned hospital admissions are due to preventable adverse drug events 3 Estimate that 1-2% inpatients harmed as a result Lewis et al (2009) Drug Safety 32:379-89 Keers et al (2013) Ann Pharmacother 47:237-56 Howard et al (2007) Br J Clin Pharmacol 63: 136-147

So what are we going to do about it?

Objectives To highlight key issues in developing, evaluating and publishing on interventions to enhance medication safety Mainly hospital practice but will also include some references to primary care

Developing interventions

Developing interventions What are the problems? Do not assume that problems (and thus solutions!) elsewhere are the same as your own Wide variation between settings and countries… Paper-based or electronic prescribing? Unit dose? Original packs? Medication preparation? Use of technology? Wide variation even within countries and settings What are the problems? - internationally a particular issue with different baseline systems for prescribing, dispensing, administering medication – different driving forces, different contexts. Eg CPOE USA vs UK Formal studies eg observations, case note reviews McLeod et al (2014). A national survey of inpatient medication systems in English NHS hospitals. BMC HSR Ahmed et al (2013). The Use and Functionality of Electronic Prescribing Systems in English Acute NHS Trusts: A Cross-Sectional Survey. PLoS ONE 8(11):

Developing interventions Focus groups Audits Incident reports Formal studies What are the problems? - internationally a particular issue with different baseline systems for prescribing, dispensing, administering medication – different driving forces, different contexts. Eg CPOE USA vs UK Formal studies eg observations, case note reviews

Developing interventions Who are the stakeholders? What are the barriers, facilitators, challenges? Plan Do Study Act (PDSA)? What are the problems? - internationally a particular issue with different baseline systems for prescribing, dispensing, administering medication – different driving forces, different contexts. Eg CPOE USA vs UK Also finance, Who are the stakeholders? Eg EP systems in hospitals – everyone thinks about the prescribers and forget about nurses (who are the biggest users!). And even if they remember about the nurses, what about the dieticians who also have to prescribe dietary supplements. And what about the patients??? What are their views on feasibility, acceptability? What are the barriers likely to be?

EVALUATING INTERVENTIONS

What are the research questions?

What are the research questions? How to increase patient safety? What are the problems? Why do they occur? What might the solutions be? What works? What works best? Which are cost-effective?

What are the research questions? How to increase patient safety? What are the problems? How often do they occur? Why do they occur? What might the solutions be? What works? What works best? Which are cost-effective?

What are the research questions? How to increase patient safety? What are the problems? How often do they occur? Why do they occur? What might the solutions be? What works? What works best? Which are cost-effective? Developing interventions Evaluating interventions

Types of question Quantitative methods Qualitative methods How many? - Clinical outcomes - Observations - Surveys - Audits Why? How? Qualitative methods - Focus groups - Interviews

Quantitative v qualitative characteristics Measuring/counting Hypothesis testing Random sampling Scientific empiricism Statistical analysis QUALITATIVE Exploring/qualifying Generates hypotheses Purposive sampling Naturalistic Eg. Content analysis, framework analysis

Quantitative methods

Quantitative methods - important issues Define what you are counting Define your denominator Choice of data collection method Validity Reliability Sampling strategy Generalisability Study design

1. Definitions Wide ranges of published error rates: Published rates of prescribing errors in England range from 1-15% of inpatient medication orders written Internationally, estimates of dispensing error rates in community pharmacy vary from 0.04% to 24% of dispensed items

1. Definitions What is, and what isn’t, an error? ?

2. Choice of data collection method Example: detection of prescribing errors in hospital Prospective reporting by pharmacists? Retrospective review of medical records & prescriptions ? Incident reports? Trigger tools?

2. How do methods compare. (n = 135 errors in total; 10 2. How do methods compare? (n = 135 errors in total; 10.7% of medication orders) Retrospective Review (n = 93; 69%) Trigger Tool (n = 0) 41 86 7 Data recorded by ward pharmacist (n = 48; 36%) 1 Incident Report (n = 1; 1%) Franklin et al. Methodological variability in detecting prescribing errors and consequences for the evaluation of interventions. Pharmacoepidemiology and Drug Safety 2009; 18: 992–999

3. Study design - what is the disadvantage of collecting data just once? Medication review intervention to reduce inpatient falls Put into place in July 56 falls logged in June Measured again in October - only 15 falls Success!!!

3. Study design - what is the disadvantage of collecting data just once? Mean July to Dec = 35 Mean Jan to June = 35

3. Study design - what is the disadvantage of collecting data just once? Time series analysis

Qualitative methods

Qualitative Methods Key principles of qualitative research Types of data: What people say they believe or do What people actually do What people actually believe The context of what people say/do/believe

mixed methods

Mixed methods Integration of qualitative and quantitative methods in the same study to answer a research question Increase in breadth and depth Various ways in which the two are integrated Independent vs interactive Equal priority vs one weighted more than the other Timing: concurrent vs sequential vs multi-phase Interface: data collection vs data analysis vs data interpretation Hadi et al (2013). Int J Pharm Prac 21: 341-45

Some examples

The Prescribing Improvement Model Study (PIMs) Improving patient safety through providing feedback to junior doctors on prescribing errors

First... identify root causes

Quotes “Also for something like aspirin, I know most pharmacists would just add that on to the drug chart and PNC [prescriber not contacted], so not contact the prescriber because it’s so small you wouldn’t contact the doctor just to say, oh it should be enteric coated or, oh it should be dispersible and you didn’t write that on..A lot of the time we’ll change, we’ll add modified release and, without probably telling the doctor”. (Pharmacist)

Quotes “And there’s another key issue here as well especially if you’re in an area where there’s a lot of doctors rotating, sometimes that phenytoin prescription is written by Doctor X, Doctor X has gone home so I have to go to Doctor Y and get them to change it and that’s fine, they learn something new, but Doctor X who wrote the prescription doesn’t know anything about it”. (Pharmacist)

Is this the problem?

Prescribing Improvement Model Aim To develop, test the feasibility, and evaluate a practical, low-cost intervention to provide feedback to junior doctors on prescribing errors and increase patient safety. Three objectives: To encourage prescribers to identify themselves when prescribing To increase the feedback given by pharmacists to individual prescribers on their prescribing errors To introduce group feedback to junior doctors on common prescribing errors Prescribing errors happen. What doesn’t always happen is that prescribers get to know about them. Part of the problem is that prescribers routinely do not identify themselves when prescribing. The pharmacists’ emphasis is often on correcting the drug chart, which only solves the problem for that particular patient. The doctor is not told when they have made an error. We wanted to change that. Mention again the toolkit as an outcome

Focus group - foundation year 1 doctors (FY1s)

And what do our patients think? “…it’s OK to screw up once but there ought to be a process that says you’ve screwed up once and we’re going to correct it so that it doesn’t happen again.  What’s unforgivable is if you’ve got the ability to go on screwing up time and time again” Patient focus group participant

1. Prescriber Identification At the three intervention sites, we gave FY1s a name stamp, a set of instructions and asked them to use it. We modified the message to identify yourself when prescribing, rather than focussing on using the name-stamp per se. We also spoke to the doctors in person, mostly on an ad-hoc basis, and asked them to identify themselves when prescribing. At imperial, we gave a short presentation to them at their weekly education sessions.

PDSA cycles Ogrinc G, Shojania KG. BMJ Qual Saf 2014;23:265–267. As a result of our PDSA cycles we added the prefix “Dr” to name-stamps, ensured we were using prescribers’ preferred names (not always the same as those held by human resources), modified our initial message from “use your name-stamp” to “state your name when prescribing”, added a label to name-stamps reminding doctors to sign their prescription, made a minor change to our inpatient drug chart and designed brief supporting information to accompany the name-stamps when distributed.

Fortnightly data Percentage of inpatient medication orders written FY1s where prescriber is identifiable

Fortnightly data Percentage of inpatient medication orders written by FY1s where prescriber is identifiable Estimate increase from 7% to 40-50%

2. Individual feedback Pharmacists asked to: Publicity and education Identify individual prescriber Contact individual prescriber Tell them an error made Suggest how to avoid the error Publicity and education Accompanied visits Building on being able to identify the prescriber Error v mistake

3. “Prescribing tips” Sent fortnightly “Spot the error” Discusses one or two errors in more depth Readable Compatible with smartphones Links to relevant prescribing resources Locally relevant Referred to specialist pharmacists during development.

Evaluation Process measures Outcome measures Weekly audit on identifiable prescribers Pharmacists assessed for feedback provision Outcome measures Prevalence of prescribing errors Questionnaire Focus groups Prescriber identification audited on a weekly basis by the project team at Imperial, and by existing pharmacists at North West London Prescribing errors audited weekly by ward pharmacists at all sites, inclusive of control and intervention and North West London Intervention and control hospitals Intervention hospital

Findings We estimate that we increased the percentage of FY1 medication orders for which the prescriber was identifiable from about 6% to 50%. Focus groups with pharmacists and FY1s suggested real benefits of our interventions and no evidence of negative unintended consequences. Attempts to produce a measureable reduction in prescribing errors are likely to need multi-faceted approach of which feedback should form part.

Hopefully...

The Dose-Reference Card (Dr-CARD) Development and evaluation of a pocket card to support prescribing by junior doctors in an English hospital

The Dr-CARD Focus groups held locally: foundation year 1 (FY1) doctors perceived time pressure and lack of access to information to be sources of stress, and to potentially contribute to erroneous prescribing. Many had developed their own pocket reference guides for commonly prescribed drugs

Dr-CARD

PUBLISHING

Publishing this work

Publishing this work Choice of journal Appropriate checklists for study design Quality improvement work SQUIRE guidelines Context – what kind of setting? Definitions What did you count as an error / adverse drug event / adverse drug reaction? Who or what was counted, and non-counted, in your denominator?

The right tools for the job