Executive Memorandum

INTRODUCTION

The disease that was most interesting to me was hypertension. I have had many issues throughout the years with high blood pressure and was curious of the data from the Garnic et al article. After reviewing the data, I was left with far more questions than answers, however. There was a total of 543 patients out of 10,842 total patients that were identified as having hypertension, in the physician’s dataset. That is only 5%. This seems very low to me. According to the CDC, approximately 1 in every 3 American adults have high blood pressure. Is this because the providers coded the disease incorrectly or is because the data was collected for the most serious diseases throughout a certain time frame. That is why the analysis of the data is vital to an organization.

---

FURTHER ANALYSIS EXPLAINED

As I broke down the data into the care quality measures by procedure I found some very interesting information. The total number of those diagnosed with hypertension were tested as follows:

• Creatinine: 68 patients
• Trig: 26 patients
• CHOL: 81 patients
• K+: 109 patients

This is a total of 284 patients receiving lab tests with a diag1 indicating hypertension. Is this because there really isn’t a specific lab test that would indicate to a provider hypertension? I would have thought that a sodium (NA+) test would be ordered given that persons with a high salt diet have a greater chance for hypertension.

Upon further analysis of the data it was identified that the sum total amount for in-patients ($15,685.37) and out-patients ($49,407.81) was $65,093.18. This did appear to be low, so further analysis was needed. The following is the average per patient:

• inp_prac: $2,240.77

All 7 in-patients

• phy_prac: $90.99

Top 12 out-patient

---

PROGRAMMER RECOMMENDATIONS

Improving patient outcomes and reducing operational costs requires a significant effort in data analysis. By focusing on the disease management program, our programmers need to extract data identifying diseases that are most prevalent and cost exorbitant. Drilling down through the layers of data is vital because a peripheral view can be misleading. We want to ensure that our providers are coding diseases accurately. By ensuring the correct coding, everyone benefits; the patients, the insurers, the providers, and the hospital/clinic.

I would recommend to the programmers to start with the most common and expensive diseases and tests. In the physician’s data set there were 4043 Laboratory examinations (V726) and the most expensive test was a Radiological exam NEC (V725), $622,968.95. Whereas in the inpatient data there were 144 patients admitted with Pneumonia, organism NOS (486) and the most expensive disease was Coronary atherosclerosis (4140), $1,511,636.77. This is the perfect place to start. Are Laboratory examinations over-ordered in the out-patient setting? Are the Radiological exams too costly? Is there something we can do for our patients to reduce Pneumonia and Coronary issues? These are the areas that I would like the programmers to identify.

---

CONCLUSION

As I stated above, I have many more questions than answers after analyzing the hypertension data. For example, were many of the patients admitted under different codes as the primary code where a secondary code could have been beneficial? When exactly was this data collected? I now the article was published in 1994, so are the ICD9 and CPT codes reflected accurately in the data? The instructions described the claims data as being from 3000+ elderly for a recent calendar year. This causes discrepancies in the analysis.

Because I am a very curious person I dug in a bit and found that 297 patients had a secondary diagnoses codes for hypertension. Additionally, there were 128 patients coded as diag3, and 42 with hypertension in the diag4 column. This would explain some of the low costs attributed to hypertension.

As an executive, asking a programmer to investigate the impact of a disease management system, I would ensure that the data is “good data”. Given this is a simulation assignment, the analysis was completed to the best of my ability. I appreciate what the programmers must go through in order to analyze data.

---

JPGs

LAB

RAD

Pneumonia

Coronary atherosclerosis

---

CODES USED

/* get count of number of distinct hypertension patients in phy file */

select count(distinct baseid) from phy_prac where diag1 = ‘40100’ or diag1 = ‘4011’ or diag1 = ‘40019’

count(distinct baseid) 543

---

select count(distinct baseid) from inp_prac where prindiag = ‘40100’ or prindiag = ‘4011’ or prindiag = ‘4019’

count(distinct baseid) 7

---

Creatinine:

select count(distinct baseid) from phy_prac where (diag1 = ‘40100’ or diag1 = ‘4011’ or diag1 = ‘40019’) and (hcpc1 between ‘80012’ and ‘80019’ or hcpc1=‘80060’or hcpc1=’80065’ or hcpc1=‘80073’ or hcpc1=‘82546’ or hcpc1=‘82565’)

count(distinct baseid) 68

---

Trig:

select count(distinct baseid) from phy_prac where (diag1 = ‘40100’ or diag1 = ‘4011’ or diag1 = ‘40019’) and (hcpc1 between ‘80061’ and ‘80062’ or hcpc1=‘80065’ or hcpc1=‘83705’ or hcpc1=‘83720’ or hcpc1=‘84478’)

count(distinct baseid) 26

---

CHOL:

select count(distinct baseid) from phy_prac where (diag1 = ‘40100’ or diag1 = ‘4011’ or diag1 = ‘40019’) and (hcpc1 between ‘80012’ and ‘80019’ or hcpc1=‘80050’ or hcpc1=‘80053’ or hcpc1 between ‘80060’ and ‘80062’ or hcpc1=‘80065’ or hcpc1=‘82465’ or hcpc1=‘82470’)

count(distinct baseid) 81

---

K+:

select count(distinct baseid) from phy_prac where (diag1 = ‘40100’ or diag1 = ‘4011’ or diag1 = ‘40019’) and (hcpc1 between ‘80002’ and ‘80019’ or hcpc1=‘80060’ or hcpc1=‘80062’ or hcpc1=‘80065’ or hcpc1=‘84132’)

count(distinct baseid) 109

---

Sum Totals

select sum(subcrg) as sumtot from phy_prac where diag1 = ‘40100’ or diag1 = ‘4011’ or diag1 = ‘40019’

sumtot 49407.81

select sum(tchgamt) as sumtot from inp_prac where prindiag = ‘40100’ or prindiag = ‘4011’ or prindiag = ‘4019’

sumtot 15685.37