HL7 Clinical Guidelines SIG Usecase Scenario: Landesh Pharmaceutical System Reform Project

HL7 Clinical Guidelines SIG Usecase Scenario: Landesh Pharmaceutical System Reform Project

Keywords

Pharmaceutical system, primary health care, clinical guidelines, prescribing decision support, evidence-based practice, best-practice, performance indicators, incentives, ICD10, ATC, ICPC, SNOMED-CT, antibiotics, upper respiratory tract infections, usecase scenario, HL7 health development framework methodology, story boards, static, active activity diagrams, term glossaries.

Country background

Landesh is a hypothetical middle income European country still recovering from recent civil war. It has moved from a central command economy to a free market system. The population is around 5 million and the per capita GDP is about US $5000. With a Balkans economy, Landesh cannot afford Swedish or US style medicine. However, the Landesh government is committed to reforming the health system in order to deliver the best possible health outcomes for the limited money available.

Primary Health Care Problems

The Landesh system of primary health care is based on capitation; primary care physicians get paid from the central health insurance agency (ZZLO) according to the number of registered patients on their list. Doctors have been concerned that the capitation payment has not kept pace with the rising costs and recently went on strike in support of higher wages. However, the health minister is concerned that merely increasing the capitation fee will encourage lazy doctors; he points to falling immunisation and Pap smear rates, increasing specialist referrals and excessive prescribing as evidence of problems with the current system.

Planned Reform

The minister is willing to increase the remuneration of primary care physicians but only if payment is linked to performance indicators that demonstrate best-practice. He wants to introduce an incentive system to encourage best-practice in which physicians would receive an additional 20% of their capitation income if they practice in accord with evidence-based, best-practice guidelines. As recommended by others, therapeutics is a pilot area to be targeted.[1]

Pharmaceutical System

Like many countries (with the notable exception of the USA), Landesh has devised their own limited list of government subsidised drugs, their own national formulary[2] and their own national drug bulletin. In addition, several groups have made a start on producing best-practice therapeutic guidelines. While much of this material is in print format, some has been converted to electronic format (HTML). There is a national information technology (IT) plan to computerise primary health care and software specifications are soon to be drawn up. However, to-date there is no agreement as to how therapeutic information should be converted into electronic formats or incorporated into the prescribing module of clinical software.

Currently, doctors (Figure 1) write prescriptions by hand but they are also obliged to add an ICD10 code to identify the clinical problem for which the drug was prescribed (otherwise the prescription cannot be dispensed). If a patient has several problems requiring the prescription of different drugs then these are required to be written on different prescription forms, each with the appropriate ICD10 code.

All pharmacies are computerised (Figure 2) and, in order to obtain reimbursement from ZZLO, pharmacists must regularly submit a floppy disk containing prescription information including ICD10 clinical diagnostic codes. This allows ZZLO to analyse prescription data according to diagnosis. It is recognised that manual ICD10 coding of prescription is not always done accurately by physicians. For example, prescriptions coded as J02 / J03 "Acute pharyngitis / tonsillitis" can contain antacids, insulin or diuretics. Regardless, an analysis of ZZLO data suggests that current prescribing practices leave considerable room for improvement (Figure 2 & 3).

Action to date

A Landesh guidelines and feedback group has prioritised 30 diseases for which guidelines should be considered (Figure 4). This list was based on data concerning the incidence of diseases in Landesh, the amount of money spent on disease specific pharmaceuticals and known prescribing problems.

Within this list, upper respiratory tract infections (common colds, tonsillitis, etc.) were selected for a pilot implementation of therapeutic guidelines because of clear evidence of inappropriate use from ZZLO data and also because excessive use of antibiotics is associated with ecological as well as economic and patient consequences.

The Landesh guidelines and feedback group then prepared local guidelines for these conditions with the help of international guideline models.[3],[4] The guidelines were then converted to a basic electronic format (HTML Help) designed for quick look-up during a clinical consultation (Figures 6 & 7).

A physician focus group has suggested that the proposed primary health care IT system should have the following characteristics with respect to integrating therapeutic guidelines, performance monitoring and best-practice incentives into the prescribing module (Figure 8):

1. Clinician selects patient (ID number or name search).
2. System displays current clinical problem list and current medication (if relevant).
3. Clinician enters patient’s clinical problem (from an ICD10[5] code pick list).
4. System asks if clinician wishes to prescribe from:

a)  specific best practice therapeutic guidelines;

b)  existing prescription in medication list (repeat prescription);

c)  writing a prescription unaided.

If a):

1. System displays best practice guideline including therapeutic recommendation.
2. Clinician can select a recommended regimen by clicking a prescribe icon alongside.
3. If selected, system drafts a prescription with patient details and the selected regimen.
4. Go to 10

If b):

1. System displays past prescriptions for that problem.

Clinician can select a past regimen by clicking a prescribe icon alongside.

If c):

1. Ultimately, the physician should also be able to access monographs in the national drug formulary (hyperlinked to therapeutic guidelines by drug names) and a drug-drug / drug-condition / drug-allergy database against information held in the patient’s electronic medical record.
2. If patient is a child, paediatric dose calculator is invoked.
3. Clinician edits prescription if necessary
4. Clinician approves (“signs”) prescription.
5. Prescription printed.
6. Prescription recorded in patient’s medication history.
7. Concordance of prescription with best-practice guidelines is tracked (using ATC codes).
8. Monthly self-audit report showing the physicians prescribing concordance with best-practice guidelines produced.
9. Six monthly reports prepared for ZZLO to enable the physician’s performance to be reviewed and an appropriate performance bonus paid).

Evaluation

In the first 12 months, the antibiotic prescribing for the common cold and tonsillitis / pharyngitis will be assessed. The following performance indicators are suggested:

ICD10 Code / Clinical problem / Prescribing performance indicators[6] / ATC Code(s) / Points
J00 / Common cold / Less than 5% of all prescriptions antibiotics:
Equal or greater than 5% of prescriptions antibiotics: / J01 / 80
0
J02 J03 / Pharyngitis / Tonsillitis / If an antibiotic prescribed;
If equal or greater than 80% narrow-spectrum[7]:
If less than 80% narrow-spectrum: / J01CE OR J01FA / 80
0

The impact of computerisation (and performance payments) will be assessed by comparing the prescribing habits of doctors using the system for the targeted conditions:

1. Each monthly basis for the first 6 months after its introduction (time series-system data).
2. Twelve months before and 12 months after its introduction using (ZZLO data).
3. To those doctors not using the system (ZZLO data).

Next steps

The above scenario needs to be refined using the HL7 Health Development Framework (HDF) methodology (story boards, static and active activity diagrams, and term glossaries).

Figures

Figure 1. Landesh physician

Figure 2. Landesh pharmacist entering prescription details into pharmacy computer.

Figure 3. HZZO GP antibiotic prescribing data for prescriptions with ICD10 codes J02 and J03 (2002).[8] There is considerable use of expensive broad-spectrum antibiotics which is not in accord with best-practice (see also Figure 6).

Figure 4. HZZO GP prescribing data for prescriptions with ICD10 codes J00 (2002).[9] There is considerable use of antibiotics which are not recommended by best-practice therapeutic guidelines (see also Figure 7).

Figure 5. Clinical problems prioritised by the Landesh
Guidelines & Feedback Advisory Group.

Figure 6. Best-practice therapeutic guideline recommendations concerning the treatment of pharyngitis / tonsillitis (ICD10 codes J03 /J03)[10]

Figure 7. Best-practice therapeutic guideline recommendations concerning the treatment of the common cold (ICD10 code J00).

Figure 8. The pharmaceutical strategy being proposed is based on the above quality assurance cycle.

Page 3

[1] Maynard A, Bloor K. Dilemmas in regulation of the market for pharmaceuticals. Health Affairs 2003; 22: 3.
http://www.healthaffairs.org/1130_abstract_c.php?ID=http://www.healthaffairs.org/Library/v22n3/s9.pdf

[2] Based on the WHO Model Formulary: http://www.who.int/medicines/organization/par/formulary.shtml

[3] http://www.ebm-guidelines.com/home.html

[4] http://www.tg.com.au

[5] ICPC or SNOMED-CT may be more suitable coding systems in the longer term

[6] It should be noted that therapeutic guidelines cannot always be prescriptive and there may be legitimate reasons for physicians to depart from them for particular patients. For this reason, 100% of the available therapeutic performance points will be allocated for less than 100% concordance with best-practice guidelines.

[7] N.B. If maximum savings are to be achieved from encouraging the prescription of narrow-spectrum antibiotics then fenoksimetilpenicillin must be added to the positive list at a price referenced close to amoksicilin.

[8] This data is provided for illustrative purposes only and does not necessarily represent that of a real country.

[9] This data is provided for illustrative purposes only and does not necessarily represent that of a real country.

.

[10] These demonstration electronic guidelines have been output in Microsoft HTML Help format but could equally have been output in a format suitable for an Intranet, the WWW or a database.