Analysis & Display of Haematology & Biochemistry Data for Oncology Patients Penny Ross University of Portsmouth & Portsmouth Hospitals Friends of Radiotherapy.

Analysis & Display of Haematology & Biochemistry Data for Oncology Patients

Penny Ross

University of Portsmouth & Portsmouth Hospitals Friends of

Radiotherapy (FORT) & Portsmouth Oncology Centre

Can the display of clinical data (Haematology & Biochemistry) be improved to provide more information to clinicians?

Does changing the presentation have an effect on clinical interpretation and decision-making?

Why is this research needed?

• The decision to treat with chemotherapy is based on a review of the patients’ blood tests when they attend clinic.

Difficult making decisions due to:• Noisy distracting workplace• Context switching, conflicting priorities• Long hours, tiredness, pressure, stress• Complex data requiring fast assimilation

Why is this research needed?

• Dangerous poisonous drugs

• Expensive

• Short shelf life for some drugs

• Many combinations and dose adjustments

• Is the patient benefiting from the chemotherapy? If not, their quality of life may improve without the drugs

• What would be the best way to present this data? What visual format is the easiest to understand?

• Line Graphs • Radar Graphs

• Colour coding • Pattern Recognition

But before designing the display there are questions about the data

Questions about the Data

• Which data should be shown together?• Is a small change in one data item more significant

than a larger change in another?• Is it the actual value of the data item that is

important or its’ change in value from the last reading?

• Are changing values that are still within normal limits important?

• Which data items should be presented together?

Study on Breast and Lung Cancer Results

• Look at the data over the period whilst chemotherapy was given

• Comparison of patients who tolerated chemotherapy well and have a good prognosis to those patients who did not benefit from chemotherapy

• Analysis of the data may reveal trends that will allow us to predict a patients performance.

Current presentation of haematology & biochemistry results only provides

information on the patients’ current state.

Plotting over the cycle shows variations.All these results are within normal limits but this shows there is a peak every day eight of the cycle

Urea

01234567

14/07/04 28/07/04 10/08/04 11/08/04 17/08/04 25/08/04 31/08/04 07/09/04 21/09/04 28/09/04

cycle 1day 1

cycle 1day 8

cycle 2 day 1 cycle 2 day 8 cycle 3day 1

cycle 3day 8

cycle 4day 1

cycle 4day 8

Urea

Again all these results are within normal limits but there is cyclic variation

Albumin

38

40

42

44

46

cycle 1 day1

cycle 1 day8

cycle 2day 1

cycle 2day 8

cycle 3day 1

cycle 3day 8

cycle 4day 1

cycle 4day 8

14/07/04 28/07/04 10/08/04 11/08/04 17/08/04 25/08/04 31/08/04 07/09/04 21/09/04 28/09/04

chemo cycle

actu

al d

ata

valu

e g

/l

Albumin

Data Analysis

Initial analysis of data based on 15 patients with either lung or metastasis or adjuvant breast cancer.

Sample selected to represent differing response to chemotherapy

Initial results show that there are trends and indications that may allow for earlier intervention

Analysis now being applied to data set of ~1,000 cases with ~7,000 sets of pathology results

Supervisors

• Dr David Prytherch

• Dr Jim Briggs

• Dr Tim Gulliford