Analysis of alternative pathway scenarios#

Aims:#

Analyse the performance of the following scenarios:

  1. Base: Uses the hospitals’ recorded pathway statistics in SSNAP (same as validation notebook)

  2. Speed: Sets 95% of patients having a scan within 4 hours of arrival, and all patients have 15 minutes arrival to scan and 15 minutes scan to needle.

  3. Onset-known: Sets the proportion of patients with a known onset time of stroke to the national upper quartile if currently less than the national upper quartile (leave any greater than the upper national quartile at their current level).

  4. Benchmark: The benchmark thrombolysis rate takes the likelihood to give thrombolysis for patients scanned within 4 hours of onset from the majority vote of the 30 hospitals with the highest predicted thrombolysis use in a standard 10k cohort set of patients. These are from Random Forests models.

  5. Combine Speed and Onset-known

  6. Combine Speed and Benchmark

  7. Combine Onset-known and Benchmark

  8. Combine Speed, Onset-known and Benchmark

The analysis will be at a global level (all hospitals together) as well as demonstrating analysis at a hospital level.

Load libraries#

import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
from scipy import stats

from matplotlib.ticker import (MultipleLocator, FormatStrFormatter)

Load data#

# Scenario results
results = pd.read_csv('./output/scenario_results.csv')
# Pathway performance paramters used in scenarios
performance_base = pd.read_csv('./output/performance_base.csv')
performance_speed = pd.read_csv('./output/performance_speed.csv')
performance_onset = pd.read_csv('./output/performance_onset.csv')
performance_speed_onset = pd.read_csv('./output/performance_speed_onset.csv')
performance_speed_benchmark = \
    pd.read_csv('./output/performance_speed_benchmark.csv')
performance_onset_benchmark = \
    pd.read_csv('./output/performance_onset_benchmark.csv')
performance_speed_onset_benchmark = \
    pd.read_csv('./output/performance_speed_onset_benchmark.csv')

performance_base
stroke_team thrombolysis_rate admissions 80_plus onset_known known_arrival_within_4hrs onset_arrival_mins_mu onset_arrival_mins_sigma scan_within_4_hrs arrival_scan_arrival_mins_mu arrival_scan_arrival_mins_sigma onset_scan_4_hrs eligable scan_needle_mins_mu scan_needle_mins_sigma
0 AGNOF1041H 0.154839 671.666667 0.425459 0.635236 0.681250 4.576874 0.557598 0.965596 1.665700 1.497966 0.935867 0.388325 3.669602 0.664462
1 AKCGO9726K 0.158892 1143.333333 0.395658 0.970845 0.428829 4.625486 0.597451 0.955882 2.834183 0.999719 0.908425 0.419355 2.904479 0.874818
2 AOBTM3098N 0.085885 500.666667 0.485470 0.619174 0.629032 4.603918 0.584882 0.935043 3.471419 1.254744 0.846435 0.267819 3.694918 0.518929
3 APXEE8191H 0.098634 439.333333 0.515679 0.716237 0.608051 4.590357 0.496452 0.966899 3.312930 0.714465 0.904505 0.258964 3.585094 0.751204
4 ATDID5461S 0.090689 275.666667 0.533546 0.573156 0.660338 4.427826 0.591373 0.878594 4.125690 0.549301 0.865455 0.315126 3.497262 0.608126
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
127 YPKYH1768F 0.105193 250.333333 0.321767 0.585885 0.720455 4.436404 0.569248 0.952681 3.779215 0.872809 0.844371 0.305882 3.982100 0.683223
128 YQMZV4284N 0.104186 358.333333 0.508511 0.945116 0.462598 4.664536 0.494740 0.948936 3.574735 0.912298 0.798206 0.308989 3.285165 0.463749
129 ZBVSO0975W 0.081602 449.333333 0.442130 0.465134 0.688995 4.562051 0.510524 0.972222 2.860226 0.990966 0.930952 0.273657 3.606046 0.575788
130 ZHCLE1578P 0.112647 796.000000 0.484694 0.733668 0.671233 4.606557 0.546648 0.949830 3.306916 0.842940 0.892569 0.262788 3.276043 0.795401
131 ZRRCV7012C 0.063058 597.333333 0.469504 0.779576 0.504653 4.636283 0.485394 0.977305 3.743456 0.662710 0.851959 0.189097 3.261270 0.803624

132 rows × 15 columns

Collate key results#

Collate key results together in a DataFrame.

# Add admission numbers to results
admissions = performance_base[['stroke_team', 'admissions']]
results = results.merge(
    admissions, how='left', left_on='stroke_team', right_on='stroke_team')

# Calculate numbers thrombolysed
results['thrombolysed'] = \
    results['admissions'] * results['Percent_Thrombolysis_(mean)'] / 100

# Calculate additional good outcomes
results['add_good_outcomes'] = (results['admissions'] * 
    results['Additional_good_outcomes_per_1000_patients_(mean)'] / 1000)

# Get key results
key_results = pd.DataFrame()
key_results['stroke_team'] = results['stroke_team']
key_results['scenario'] = results['scenario']
key_results['admissions'] = results['admissions']
key_results['thrombolysis_rate'] = results['Percent_Thrombolysis_(mean)']
key_results['additional_good_outcomes_per_1000_patients'] = \
    results['Additional_good_outcomes_per_1000_patients_(mean)']
key_results['patients_receiving_thrombolysis'] = results['thrombolysed']
key_results['add_good_outcomes'] = results['add_good_outcomes']

# Remove same_patient_characteristics as not needed here
mask = key_results['scenario'] != 'same_patient_characteristics'
key_results = key_results[mask]

key_results
stroke_team scenario admissions thrombolysis_rate additional_good_outcomes_per_1000_patients patients_receiving_thrombolysis add_good_outcomes
0 AGNOF1041H base 671.666667 15.20 12.76 102.093333 8.570467
1 AKCGO9726K base 1143.333333 14.91 13.21 170.471000 15.103433
2 AOBTM3098N base 500.666667 7.80 5.67 39.052000 2.838780
3 APXEE8191H base 439.333333 10.40 7.59 45.690667 3.334540
4 ATDID5461S base 275.666667 9.17 6.28 25.278633 1.731187
... ... ... ... ... ... ... ...
1051 YPKYH1768F speed_onset_benchmark 250.333333 22.20 23.54 55.574000 5.892847
1052 YQMZV4284N speed_onset_benchmark 358.333333 13.51 12.02 48.410833 4.307167
1053 ZBVSO0975W speed_onset_benchmark 449.333333 21.12 20.10 94.899200 9.031600
1054 ZHCLE1578P speed_onset_benchmark 796.000000 13.89 12.82 110.564400 10.204720
1055 ZRRCV7012C speed_onset_benchmark 597.333333 12.25 11.22 73.173333 6.702080

1056 rows × 7 columns

key_results.to_csv('./output/key_scenario_results.csv', index=False)

Overall results#

columns = ['admissions', 'patients_receiving_thrombolysis', 'add_good_outcomes']
summary_stats = key_results.groupby('scenario')[columns].sum()
summary_stats['percent_thrombolysis'] = (100 *
    summary_stats['patients_receiving_thrombolysis'] / summary_stats['admissions'])
summary_stats['add_good_outcomes_per_1000'] = (1000 *
    summary_stats['add_good_outcomes'] / summary_stats['admissions'])

# Re-order
order = {'base': 1, 'speed': 2, 'onset': 3, 'benchmark': 4, 'speed_onset': 5,
    'speed_benchmark': 6, 'onset_benchmark':7, 'speed_onset_benchmark': 8,
    'same_patient_characteristics': 9}
df_order = [order[x] for x in list(summary_stats.index)]
summary_stats['order'] = df_order
summary_stats.sort_values('order', inplace=True)

# Select cols of interest
summary_stats = summary_stats[['percent_thrombolysis', 'add_good_outcomes_per_1000']]

base_thrombolysis = summary_stats.loc['base']['percent_thrombolysis']
summary_stats['Percent increase thrombolysis'] = (100 * (
        summary_stats['percent_thrombolysis'] / base_thrombolysis -1))

base_add_good_outcomes = summary_stats.loc['base']['add_good_outcomes_per_1000']
summary_stats['Percent increase good_outcomes'] = (100 * (
    summary_stats['add_good_outcomes_per_1000'] / base_add_good_outcomes -1))

summary_stats = summary_stats.round(2)

summary_stats.to_csv('./output/summary_net_results.csv')

summary_stats
percent_thrombolysis add_good_outcomes_per_1000 Percent increase thrombolysis Percent increase good_outcomes
scenario
base 11.60 9.40 0.00 0.00
speed 12.34 11.86 6.37 26.13
onset 13.67 11.09 17.86 17.87
benchmark 14.51 11.70 25.05 24.36
speed_onset 14.52 13.98 25.16 48.63
speed_benchmark 15.43 14.82 33.01 57.59
onset_benchmark 17.20 13.87 48.32 47.53
speed_onset_benchmark 18.27 17.57 57.54 86.82 
fig = plt.figure(figsize=(10,7))

ax1 = fig.add_subplot(121)
x = list(summary_stats.index)
y1 = summary_stats['percent_thrombolysis'].values
ax1.bar(x,y1)
ax1.set_ylim(0,20)
plt.xticks(rotation=90)
plt.yticks(np.arange(0,22,2))
ax1.set_title('Thrombolysis use (%)')
ax1.set_ylabel('Thrombolysis use (%)')
ax1.set_xlabel('Scenario')
ax1.grid(axis = 'y')

ax2 = fig.add_subplot(122)
x = list(summary_stats.index)
y1 = summary_stats['add_good_outcomes_per_1000'].values
ax2.bar(x,y1, color='r')
ax2.set_ylim(0,20)
plt.xticks(rotation=90)
plt.yticks(np.arange(0,22,2))
ax2.set_title('Additional good outcomes\nper 1,000 admissions')
ax2.set_ylabel('Additional good outcomes\nper 1,000 admissions')
ax2.set_xlabel('Scenario')
ax2.grid(axis = 'y')

plt.tight_layout(pad=2)

plt.savefig('./output/global_change.jpg', dpi=300)

plt.show()
../_images/scenario_analysis_13_0.png

Plot changes at each hospital#

def compare_plot(base_rx, test_rx, base_benefit, test_benefit, name):
    
    # Set up sublot
    fig, ax = plt.subplots(1,2, figsize=(10,6))
    
    # Thrombolysis use    
    x = base_rx
    y = test_rx
    for i in range(len(x)):
        if y[i] >= x[i]:
            ax[0].plot([x[i],x[i]],[x[i],y[i]],'g-o', alpha=0.6)
        else:
            ax[0].plot([x[i],x[i]],[x[i],y[i]],'r-o', alpha=0.6)
    ax[0].set_xlim(0)
    ax[0].set_ylim(0)
    ax[0].grid()
    ax[0].set_xlabel('Base thrombolysis use (%)')
    ax[0].set_ylabel('Scenario thrombolysis use (%)')
    ax[0].set_title('Thrombolysis use')
    
    # Benefit
    x = base_benefit
    y = test_benefit
    for i in range(len(x)):
        if y[i] >= x[i]:
            ax[1].plot([x[i],x[i]],[x[i],y[i]],'g-o', alpha=0.6)
        else:
            ax[1].plot([x[i],x[i]],[x[i],y[i]],'r-o', alpha=0.6)
    ax[1].set_xlim(0)
    ax[1].set_ylim(0)
    ax[1].grid()
    ax[1].set_xlabel('Base benefit')
    ax[1].set_ylabel(f'Scenario benefit')
    ax[1].set_title('Clinical benefit\n(additional good outcomes for 1,000 admissions)')
    
    
    # Make axes places consistent
    ax[0].xaxis.set_major_locator(MultipleLocator(5))
    ax[0].yaxis.set_major_locator(MultipleLocator(5))
    ax[1].xaxis.set_major_locator(MultipleLocator(5))
    ax[1].yaxis.set_major_locator(MultipleLocator(5))
    ax[0].xaxis.set_major_formatter(FormatStrFormatter('%.0f'))
    ax[0].yaxis.set_major_formatter(FormatStrFormatter('%.0f'))
    ax[1].xaxis.set_major_formatter(FormatStrFormatter('%.0f'))
    ax[1].yaxis.set_major_formatter(FormatStrFormatter('%.0f'))
    
    fig.tight_layout(pad=2)
    plt.savefig(f'{name}.jpg', dpi=300)

Plot effect of speed#

base_rx = key_results[key_results['scenario'] == 'base']['thrombolysis_rate'].values
scenario_rx = key_results[key_results['scenario'] == 'speed']['thrombolysis_rate'].values
base_benfit = key_results[key_results['scenario'] == 'base']['additional_good_outcomes_per_1000_patients'].values
scenario_benefit = key_results[key_results['scenario'] == 'speed']['additional_good_outcomes_per_1000_patients'].values


compare_plot(base_rx, scenario_rx, base_benfit, scenario_benefit,
                     './output/hospitals_speed')
../_images/scenario_analysis_17_0.png

Plot effect of onset-known#

base_rx = key_results[key_results['scenario'] == 'base']['thrombolysis_rate'].values
scenario_rx = key_results[key_results['scenario'] == 'onset']['thrombolysis_rate'].values
base_benfit = key_results[key_results['scenario'] == 'base']['additional_good_outcomes_per_1000_patients'].values
scenario_benefit = key_results[key_results['scenario'] == 'onset']['additional_good_outcomes_per_1000_patients'].values


compare_plot(base_rx, scenario_rx, base_benfit, scenario_benefit,
                     './output/hospitals_onset')
../_images/scenario_analysis_19_0.png

Plot effect of benchmark decisions#

base_rx = key_results[key_results['scenario'] == 'base']['thrombolysis_rate'].values
scenario_rx = key_results[key_results['scenario'] == 'benchmark']['thrombolysis_rate'].values
base_benfit = key_results[key_results['scenario'] == 'base']['additional_good_outcomes_per_1000_patients'].values
scenario_benefit = key_results[key_results['scenario'] == 'benchmark']['additional_good_outcomes_per_1000_patients'].values


compare_plot(base_rx, scenario_rx, base_benfit, scenario_benefit,
                     './output/hospitals_benchmark')
../_images/scenario_analysis_21_0.png

Plot effect of all changes#

base_rx = key_results[key_results['scenario'] == 'base']['thrombolysis_rate'].values
scenario_rx = key_results[key_results['scenario'] == 'speed_onset_benchmark']['thrombolysis_rate'].values
base_benfit = key_results[key_results['scenario'] == 'base']['additional_good_outcomes_per_1000_patients'].values
scenario_benefit = key_results[key_results['scenario'] == 'speed_onset_benchmark']['additional_good_outcomes_per_1000_patients'].values


compare_plot(base_rx, scenario_rx, base_benfit, scenario_benefit,
                     './output/hospitals_speed_onset_benchmark')
../_images/scenario_analysis_23_0.png

Histogram of shift in distribution of thrombolysis use and benefit#

base_rx = key_results[key_results['scenario'] == 'base']['thrombolysis_rate'].values
scenario_rx = key_results[key_results['scenario'] == 'speed_onset_benchmark']['thrombolysis_rate'].values
base_benfit = key_results[key_results['scenario'] == 'base']['additional_good_outcomes_per_1000_patients'].values
scenario_benefit = key_results[key_results['scenario'] == 'speed_onset_benchmark']['additional_good_outcomes_per_1000_patients'].values


fig, ax = plt.subplots(1,2,figsize=(10,6))

bins = np.arange(1,32)
ax[0].hist(base_rx, bins = bins, color='k', linewidth=2,
           label='control', histtype='step')
ax[0].hist(scenario_rx, bins = bins, color='0.7', linewidth=2,
           label='all in-hospital changes', histtype='stepfilled')
ax[0].grid()
ax[0].set_xlabel('Thrombolysis use (%)')
ax[0].set_ylabel('Hospital count')
ax[0].set_ylim(0,30)
ax[0].yaxis.set_major_locator(MultipleLocator(5))
ax[0].set_title('Thrombolysis use')
ax[0].legend()

ax[1].hist(base_benfit, bins = bins, color='k', linewidth=2,
           label='control', histtype='step')
ax[1].hist(scenario_benefit, bins = bins, color='0.7', linewidth=2,
           label='all in-hospital changes', histtype='stepfilled')
ax[1].grid()
ax[1].set_xlabel('Clincial benefit\n(additional good outcomes per 1000 admissions)')
ax[1].set_ylabel('Hospital count')
ax[1].set_ylim(0,26)
ax[1].yaxis.set_major_locator(MultipleLocator(5))
ax[1].set_title('Clincial benefit')
ax[1].legend()

plt.savefig('./output/histograms.jpg', dpi=300)
../_images/scenario_analysis_25_0.png

Find which change makes most difference in each hospital#

Best thrombolysis rate#

# Pivot results
results_pivot = (key_results[['thrombolysis_rate', 'scenario', 'stroke_team']].pivot(
    index='stroke_team', columns='scenario'))
results_pivot = results_pivot['thrombolysis_rate']
# Limit to single changes
cols_to_drop = ['speed_benchmark', 'speed_onset', 'onset_benchmark', 'speed_onset_benchmark']
results_pivot.drop(cols_to_drop, axis=1, inplace=True)
# Find largest effect
best_result = results_pivot.idxmax(1)
# Count
best_result.value_counts()

benchmark    83
onset        39
speed        10
dtype: int64

Best outcomes#

# Pivot results
results_pivot = (key_results[['additional_good_outcomes_per_1000_patients', 
        'scenario', 'stroke_team']].pivot(index='stroke_team', columns='scenario'))
results_pivot = results_pivot['additional_good_outcomes_per_1000_patients']
# Limit to single changes
cols_to_drop = ['speed_benchmark', 'speed_onset', 'onset_benchmark', 'speed_onset_benchmark']
results_pivot.drop(cols_to_drop, axis=1, inplace=True)
# Find largest effect
best_result = results_pivot.idxmax(1)
# Count
best_result.value_counts()

benchmark    56
speed        49
onset        27
dtype: int64

Bar charts of individual changes at all hospitals#

Here we summarise the effect of individual changes at each hospital.

Note that actual combined changes will be more than additive, but these plots give an indication of what the most significant effects will be across all hospitals.

# Pivot results by scenario type
results_pivot = key_results.pivot(index='stroke_team', columns='scenario')

hosp_per_chart = np.ceil(results_pivot.shape[0]/2)

# Thrombolysis chart
fig, axs = plt.subplots(2,1, figsize=(12,12), sharey=True)
# 4 subplots
i=0
for ax in axs.flat:
    # Get subgroup of data for plot
    start = int(hosp_per_chart * i)
    end = int(hosp_per_chart * (i + 1))
    subgroup = results_pivot.iloc[start:end]
    # Get effect of speed (avoid negatives)
    speed = subgroup['thrombolysis_rate']['speed'] - subgroup['thrombolysis_rate']['base']
    speed = list(map (lambda x: max(0,x), speed))
    # Get effect of known onset (avoid negatives)
    onset = subgroup['thrombolysis_rate']['onset'] - subgroup['thrombolysis_rate']['base']
    onset = list(map (lambda x: max(0,x), onset))
    # Get effect of decision (avoid negatives)
    eligible = subgroup['thrombolysis_rate']['benchmark'] - subgroup['thrombolysis_rate']['base']
    eligible = list(map (lambda x: max(0,x), eligible))
    
    x = range(start, start + subgroup.shape[0])
    ax.bar(x, speed, color='b', label = 'Speed')
    ax.bar(x, onset, color='g', bottom = speed, label = 'Onset known')
    ax.bar(x, eligible, color='r', bottom = np.array(speed) + np.array(onset),
           label = 'Benchmark decision')
    ax.legend()
    ax.set_xlabel('Hospital')
    ax.set_ylabel('Increase in thrombolysis use (percentage points)')
    # Put y tick label son all charts
    ax.yaxis.set_tick_params(which='both', labelbottom=True)
    
    i += 1
      
plt.tight_layout(pad=2)
plt.savefig('./output/all_hosp_bar_thrombolysis.jpg', dpi=300)
plt.show()
../_images/scenario_analysis_33_0.png 
hosp_per_chart = np.ceil(results_pivot.shape[0]/2)

# Outcomes chart
fig, axs = plt.subplots(2,1, figsize=(12,12), sharey=True)
# 4 subplots
i=0
for ax in axs.flat:
    # Get subgroup of data for plot
    start = int(hosp_per_chart * i)
    end = int(hosp_per_chart * (i + 1))
    subgroup = results_pivot.iloc[start:end]
    
    # Get effect of speed (avoid negatives)
    speed = subgroup['additional_good_outcomes_per_1000_patients']['speed'] - \
        subgroup['additional_good_outcomes_per_1000_patients']['base']
    speed = list(map (lambda x: max(0,x), speed))
    # Get effect of known onset (avoid negatives)
    onset = subgroup['additional_good_outcomes_per_1000_patients']['onset'] - \
        subgroup['additional_good_outcomes_per_1000_patients']['base']
    onset = list(map (lambda x: max(0,x), onset))
    # Get effect of decision (avoid negatives)
    eligible = subgroup['additional_good_outcomes_per_1000_patients']['benchmark'] - \
        subgroup['additional_good_outcomes_per_1000_patients']['base']
    eligible = list(map (lambda x: max(0,x), eligible))
    
    x = range(start, start + subgroup.shape[0])
    ax.bar(x, speed, color='b', label = 'Speed')
    ax.bar(x, onset, color='g', bottom = speed, label = 'Onset known')
    ax.bar(x, eligible, color='r', bottom = np.array(speed) + np.array(onset),
           label = 'Benchmark decision')
    ax.legend()
    ax.set_xlabel('Hospital')
    ax.set_ylabel('Increase in good outcomes (per 1,000 admissions)')
    # Put y tick label son all charts
    ax.yaxis.set_tick_params(which='both', labelbottom=True)
    
    i += 1
      
plt.tight_layout(pad=2)
plt.savefig('./output/all_hosp_bar_outcomes.jpg', dpi=300)
plt.show()
../_images/scenario_analysis_34_0.png

Results for individual hospitals#

We may plot more detailed results at an individual hospital level.

def plot_hospital(data, id):
    
    hospital_data = data.iloc[id]
    
    max_val = max(hospital_data['thrombolysis_rate'].max(),
                  hospital_data['additional_good_outcomes_per_1000_patients'].max())
    
    max_val = 5 * int(max_val/5) + 5
    
    team = hospital_data.name
    
    # Sort results
    
    df = pd.DataFrame()
    df['thrombolysis_rate'] = hospital_data['thrombolysis_rate']
    df['outcomes'] = hospital_data['additional_good_outcomes_per_1000_patients']
    order = {'base': 1, 'speed': 2, 'onset': 3, 'benchmark': 4, 'speed_onset': 5,
    'speed_benchmark': 6, 'onset_benchmark':7, 'speed_onset_benchmark': 8}
    df_order = [order[x] for x in list(df.index)]
    df['order'] = df_order
    df.sort_values('order', inplace=True)
    
    fig = plt.figure(figsize=(10,7))

    ax1 = fig.add_subplot(121)
    x = df['thrombolysis_rate'].index
    y1 = df['thrombolysis_rate']
    ax1.bar(x,y1)
    plt.xticks(rotation=90)
    ax1.set_title('Thrombolysis use (%)')
    ax1.set_ylabel('Thrombolysis use (%)')
    ax1.set_xlabel('Scenario')
    ax1.set_ylim(0, max_val)
    ax1.grid(axis = 'y')

    ax2 = fig.add_subplot(122)
    y1 = df['outcomes']
    ax2.bar(x,y1, color='r')
    plt.xticks(rotation=90)
    ax2.set_title('Additional good outcomes\nper 1,000 admissions')
    ax2.set_ylabel('Additional good outcomes\nper 1,000 admissions')
    ax2.set_xlabel('Scenario')
    ax2.set_ylim(0, max_val)
    ax2.grid(axis = 'y')
    
    plt.suptitle(f'Scenario results for team: {team}')

    plt.tight_layout(pad=2)
    
    plt.savefig(f'./output/hosp_results_{team}.jpg', dpi=300)
    
    plt.show()

An example where speed makes most difference.

plot_hospital(results_pivot, 103)
../_images/scenario_analysis_38_0.png

An example where determining stroke onset time makes most difference.

plot_hospital(results_pivot, 64)
../_images/scenario_analysis_40_0.png

An example where applying benchmark decion-making makes most difference.

plot_hospital(results_pivot, 54)
../_images/scenario_analysis_42_0.png

Observations#

  • The combined effect of the changes modelled were to increase thrombolysis use from 11.6% to 18.3% of all patients (57% increase), and to increase the number of thrombolysis-related good outcomes per 1,000 admissions from 9.4 to 17.6 (86% increase).

  • Overall the net improvements made to thrombolysis use were benchmark decisions > determining stroke onset time > speed improvement.

  • Overall the net improvements made to additional good outcomes were speed improvement > benchmark decisions > determining stroke onset time.

  • Speed improvement therefore had least effect on the proportion of patients receiving thrombolysis, but most effect on the the predicted outcomes. This is due to improving the speed of thrombolysis improving the outcomes for all patients, including those that would have received thrombolysis anyway (without speed improvement).

  • When asking the question “which single change makes most difference at each hospital?” then for thrombolysis the order is: benchmark decisions (83/132 hospitals), determining time of stroke onset (39/132), speed (10/132). For improvements in clinical outcome the order is benchmark decisions (56/132 hospitals), speed (49/132), and determining time of stroke onset (27/132).

  • Combining improvements to the pathway had a greater effect than any single change alone.

  • Even with all improvements, there would still be expected to be a significant inter-hospital variation in use of thrombolysis and, the benefit gained. With the modelled improvements there is a general improvement across hospitals, and not a significant reduction in inter-hospital variation. The remaining variation is due to variation in patient populations.

  • Charts may be readily produced for each hospitals showing the potential benefit of each process change individually or combined.