Have a personal or library account? Click to login
Integrating the Principles of Evidence Based Medicine and Evidence Based Public Health: Impact on the Quality of Patient Care and Hospital Readmission Rates in Jordan Cover

Integrating the Principles of Evidence Based Medicine and Evidence Based Public Health: Impact on the Quality of Patient Care and Hospital Readmission Rates in Jordan

Open Access
|Aug 2016

Figures & Tables

ijic-16-3-2436-g1.png
Figure 1

Conceptual Framework.

Table 1

Multiple linear regression results for all readmissions and avoidable readmissions.

All ReadmissionsAvoidable Readmissions
EstimateSignificanceEstimateSignificance
(Intercept)2.8050.0292.4430.277
Age–0.0450.027–0.0450.016
Gender=Male–0.2120.516–1.4440.104
Marital Status=Single0.8590.1732.4940.117
Insurance Status= ≥20%–0.0670.1540.8530.181
Distance in Km=’> 60’5.6300.2026.3140.151
Distance in Km=’15–29’–0.7260.419–1.9860.140
Distance in Km=’30–44’0.0300.965–0.3920.739
Distance in Km=’45–59’0.7550.2211.6120.238
Number of Comorbidities–0.0290.8060.0320.553
Number of Admissions in the Last Year0.0730.0370.1050.005
Smoking Status=Smoker0.9480.1570.6090.574
Discharged Against Medical Advice=No0.4110.4770.3780.625
Compliant with Treatment=Not Compliant–0.3250.629–0.7340.387
Discharge Instructions= Incomplete0.5140.2521.6000.072
Follow-Up-Plan=Unclear0.2160.8280.0110.996
Discharge during Weekends or Weekdays=Weekends–0.1520.765–0.4490.658
Number Of Procedures0.7490.0000.7210.000
Delay (Time between Discharge and Readmission)0.0070.722–0.0060.853
Table 2

All admission and readmission factors with significant t-test results.

1- All Admissions N=3962T-Statistic Values (–/+) and Significance Level (*/**)
H0 : μ1μ2= 0
Regarding groups of
LOSAgeDelayAdmissions-Year back# Procedures# Comorbidity
Gender (Male – Female)–1.7700.763–0.997–0.7151.016– 0.531
Smoking (Non – Yes)1.300–4.089**0.034–8.626**0.307–11.258**
Compliant with treatment (No – Yes)1.3310.871–0.347–1.3931.7600.066
Marital Status (Single – Married)1.281–43.564**0.2410.428–4.247**–18.468**
Discharge (Weekend – Weekday)0.480–2.529*1.632–0.465– 0.040– 0.930
Coinsurance (<20% – >20%)2.449–9.360**–0.11315.699**–1.7330.016
Discharge Decision (Not AMA – Against Medical)–0.199–0.8420.315–1.141–1.821–2.645
Follow-up plan (Unclear – Clear)1.2002.685–0.9931.0631.4073.176*
Discharge Instructions (Unclear – Clear)0.6000.664–1.2251.388–0.1380.625
 
2- Readmissions N=1157T-Statistic Values (–/+) and Significance Level (*/**)
H0 : μ1μ2= 0
Regarding groups of
LOSAgeDelayAdmissions-Year back# Procedures# Comorbidity
Gender (Male – Female)–1.1401.250–0.997–1.514–0.023–2.175*
Smoking (Non – Yes)0.157–3.798**0.034–4.394**–0.045–5.611**
Compliant with Treatment (No – Yes)1.2241.125–0.866–0.9082.550*–0.211
Marital Status (Single – Married)1.921–24.721**0.2410.034–1.006–11.279**
Discharge (Weekend – Weekday)–0.216–2.3071.6320.579–0.3810.779
Coinsurance (<20%–>20%)1.264–7.166**–0.1137.695**–2.862**–1.306
Discharge Decision (Not AMA – Against Medical)–1.148–1.2510.3151.064–2.890*–1.361
Follow-up plan (Clear –Unclear)0.7581.259–0.993–2.083*0.9120.589
Discharge Instructions (Clear – Unclear)1.4540.807–1.225–1.6141.313– 0.252

[i] (*): Significant at 0.05, (**): Significant at 0.001.

Table 3

Avoidable readmission factors with significant t-test results.

Avoidable readmissions     N=506T-Statistic Values (–/+) and Significance Level (*/**)
H0 : μ1μ2= 0
Regarding groups of
LOSAgeDelayAdmissions-Year back# Procedures# Comorbidity
Gender (Male – Female)–2.420*–0.466–0.9271.463–1.094–0.999
Smoking (Non – Yes)0.113–1.2590.557–1.8450.603–2.544*
Compliant with Treatment (No – Yes)0.388–0.358–0.778–0.1391.339–1.738
Marital Status (Single – Married)2.461*–17.935**0.1412.298 *–0.705–8.047**
Discharge (Weekend – Other)–0.212–2.7791.4460.606–0.232–0.585
Coinsurance (<20%–>20%)1.256–5.146**–0.9294.900**–0.978–0.640
Discharge Decision (Not AMA – Against Medical)–0.7810.2860.183–0.568–1.899– 0.228
Follow-up plan (Clear – Unclear)–0.8360.633–0.238–1.379– 0.795–1.125
Discharge Instructions (Clear – Unclear)0.525–1.247–0.377–1.383–0.972–3.640**

[i] (*): Significant at 0.05, (**): Significant at 0.001.

Table 4

JRip data mining rules for avoidable readmissions.

Rules for all readmission (readmissions vs. no readmissions)
ROC area= 0.71 and Accuracy > 75%
Predicted classAccuracy
Rule 1: (MAIN_DISEASE_CAT = ICD-10 CAT 2) and (ADMISIONS_AYEAR_BACK >= 5)Readmission74%
Rule 2: (MAIN_DISEASE_CAT = ICD-10 CAT 2)Readmission69%
Rule 3: (ADMISIONS_AYEAR_BACK >= 2, (DISCHARGE_INSTRUCTIONS = no evidence of clear), and (AGE <= 65)Readmission73%
Rule 4: (ADMISIONS_AYEAR_BACK >= 3) and (DISCHARGE_DECISION = against medical advice)Readmission77%
Rule 5: (COINSURANCE = <20), (FOLLOW_UP_PLAN = no evidence of clear), (MED_COMPLIANT = not compliant with treatment), and (ADMISIONS_AYEAR_BACK >= 1)Readmission94%
Rule 6: (ADMISIONS_AYEAR_BACK >= 2), (DISCHARGE_INSTRUCTIONS = no evidence of clear), (NUMBER_OF_PROCEDURES <= 1), (NUM_COMORBIDITY <= 3), and (AGE >= 33)Readmission82%
 
Rules of potentially avoidable readmissions (avoidable vs. unavoidable)Predicted classAccuracy
Rule 1: (AGE >= 61), (FOLLOW_UP_PLAN = no evidence of clear), and (MAIN_DISEASE_CAT = ICD-10 CAT 10)Avoidable83%
Rule 2: (AGE >= 61) and (DISCHARGE_DECISION = against medical advice)Avoidable89%
Rule 3: (MED_COMPLIANT = not compliant)Avoidable100%
Rule 4: (FOLLOW_UP_PLAN = no evidence of clear), (AGE >= 56), (NUMBER_OF_PROCEDURES >= 9) and (NUMBER_OF_PROCEDURES <= 14)Avoidable84%
Rule 5: (FOLLOW_UP_PLAN = no evidence of clear), (AGE >= 61), (NUMBER_OF_PROCEDURES <= 4), (DELAY <= 10), and (LOS >= 3)Avoidable84%
Rule 6: (DISCHARGE_INSTRUCTIONS = no evidence of clear), (DELAY > 12), (DELAY < 16), and (NUM_COMORBIDITY >= 3)Avoidable75%
Rule 7: (FOLLOW_UP_PLAN = no evidence of clear), (SMOKING STATUS = smoker), and (ADMISIONS_AYEAR_BACK <= 3)Avoidable79%
Rule 8: (DISCHARGE_DECISION = against medical advice)Avoidable86%
Rule 9: (FOLLOW_UP_PLAN = no evidence of clear), (MAIN_DISEASE_CAT = ICD-10 CAT 5), and (AGE <= 57)Avoidable79%

[i] Abbreviations:

ICD-10 CAT 2: Malignant Neoplasms.

ICD-10 CAT 5: Endocrine, nutritional and metabolic diseases.

ICD-10 CAT 10: Diseases of the circulatory system.

Table 5

Logistic regression model of factors influencing all readmissions & avoidable readmissions.

All ReadmissionsAvoidable Readmission
FactorLevelsAdjusted Odd’s Ratio95% Confidence IntervalAdjusted Odd’s Ratio95% Confidence Interval
GenderFemale
Male
0.9710.847–1.1140.8790.729–1.061
SmokingYes
No
2.3181.953–.751*1.8161.393–2.367*
CompliantYes
No
0.0540.0270.109*0.7600.470–1.230
Marital statusMarried
Single
0.9810.815–1.1810.9630.749–1.238
DischargeWeekday
Weekend
1.0370.827–1.3001.3170.943–1.838
Coinsurance≥20%
<20%
0.3930.032–.047*1.6171.145–2.282*
Discharge DecisionNot AMA
AMA
0.2470.1750.349*0.1060.057–0.197*
Follow-up planClear
Unclear
0.6620.565–0.776*0.3380.249–0.460*
Discharge InstructionClear
Unclear
0.6330.565–0.779*0.3370.277–0.512*
Age1.0960.994–1.111*1.0191.012–1.025*
LOS1.1661.143–1.189*1.0271.008–1.046*
DelayNA0.9980.982–1.013
Adm. Year Back1.0221.010–1.034*0.9630.939–0.989*
Number of procedures0.9920.981–1.0031.0431.021–1.065*
Number of co morbidity1.1091.069–1.150*1.1961.124–1.272*

[i] (*): Significant at 0.05.

ijic-16-3-2436-g2.png
Figure 2

Heuristic guide impacting hospital readmissions.

DOI: https://doi.org/10.5334/ijic.2436 | Journal eISSN: 1568-4156
Language: English
Submitted on: Jan 15, 2016
Accepted on: Aug 16, 2016
Published on: Aug 31, 2016
Published by: Ubiquity Press
In partnership with: Paradigm Publishing Services
Publication frequency: 4 issues per year

© 2016 Mohammad S. Alyahya, Heba H. Hijazi, Hussam A. Alshraideh, Mohammad Aser Alsharman, Rabah Al Abdi, Heather Lea Harvey, published by Ubiquity Press
This work is licensed under the Creative Commons Attribution 4.0 License.