Reaction Paper

3.pdf

Home >Nursing homework help >Reaction Paper

The Effect of Ordinances Requiring Smoke-Free Restaurants and Bars on Revenues: A Follow-Up

A B S T R A C T

Objectives. The purpose of this study was to extend an earlier evalua- tion of the economic effects of ordinances requiring smoke-free res- taurants and bars.

Methods. Sales tax data for 15 cities with smoke-free restaurant ordi- nances, 5 cities and 2 counties with smoke-free bar ordinances, and matched comparison locations were analyzed by multiple regression, in- cluding time and a dummy variable for the ordinance.

Re.sults. Ordinances had no sig- nificant effect on the fraction of total retail sales that went to eating and drinking places or on the ratio between sales in communities with ordinances and sales in comparison communities. Ordinances requiring smoke-free bars had no significant effect on the fraction of revenues going to eating and drinking places that serve all types of liquor.

Conclusions. Smoke-free ordi- nances do not adversely affect either restaurant or bar sales. {Am J Public Health. 1997;87:1687-1693)

Stanton A. Glantz, PhD, and Lisa R. A. Smith

Introduction

By March 1997, more than 150 communities in the United States had eliminated smoking in public places and workplaces.' Califomia law now requires that all restaurants be smoke-free and that all bars become smoke-free on January 1, 1998.2 In 1994, using sales tax data, we evaluated the effects of ordinances requir- ing smoke-free restaurants on restaurant revenues in the first 15 US cities that had passed ordinances prohibiting smoking in the enclosed areas of restaurants (not necessarily including bar areas).^ We found that restaurant revenues were not affected, and subsequent studies con- firmed this result.'*^ The tobacco industry and its front groups continue to claim that these ordinances create severe economic problems for restaurants and bars.^"" We have added 3 more years of data (through the second quarter of 1996) for the original 15 cities (Tahle 1) as well as data for the first 5 cities and 2 counties to require bars to be smoke-free (Table 2).

Methods

We conducted two sets of analyses: (1) an analysis of the effects of smoke-free restaurant ordinances on restaurant rev- enues, and (2) an analysis of the effects of smoke-free bar ordinances on bar rev- enues.

As before, we obtained data on taxable restaurant sales and total retail sales'2-'3 for communities that had smoke- free restaurant ordinances in force as well as for comparison cities (matched on population, income, smoking prevalence, and geographic location) that provided less than 60% of seating for nonsmokers (Table 1; Point Arena is the comparison city for Ross, because Tiburon, the comparison city in our earlier study, passed a smoke-free ordinance). Analysis of the restaurants in Califomia compari- son cities could not go beyond December 1994, after which a state law required all

restaurants to be smoke-free.^ All compari- son cities were selected before the statisti- cal analysis was performed.

In the study of smoke-free bar ordinances, all communities with ordi- nances that clearly identified bars as smoke-free and that had been in effect long enough for us to obtain 1 year of sales tax data were included. For the five cities that require bars to be smoke-free, sales tax data were obtained from the Research and Statistics Division of the Califomia Board of Equalization for 1991 through 1995. Data for the two counties with smoke-free bar ordinances, pub- lished in quarterly reports,'^ were avail- able from 1986 through 1996. So that we could examine effects on eating and drinking establishments that sell all types of liquor, we gathered sales tax data separately for eating places serving no alcoholic beverages (category 24), those serving beer and wine (category 35), and those serving all types of liquor (category 36). Detailed breakdowns of revenues hy category number are available only for 1991 and later years; the Board of Equalization had disposed of the detailed data from earlier years. For San Luis Obispo, however, we have these data from an earlier study.'"* No comparison city is available for San Luis Obispo because its ordinance went into effect in August 1990. Data were also obtained for comparison communities without smoke-free bar ordi- nances (Table 2).

The analysis was conducted as be- fore. Briefly, we computed (1) the fraction of total retail sales at restaurants and (2) the ratio of restaurant sales in cities with ordinances to restaurant sales in compa- rable cities without ordinances. The linear regression analysis included time; a dummy variable, L, that indicated whether

The authors are with the Institute for Health Policy Studies, Department of Medicine, Univer- sity of Califomia, San Francisco; Dr Glantz is also with the Division of Cardiology.

Requests for reprints should be sent to Stanton A, Glantz, PhD. Division of Cardiology. Box 0124, San Francisco, CA 94143-0124,

This paper was accepted August 8, 1997,

October 1997, Vol. 87. No, 10 American Joumal of Public Health 1687

Public Health Briefs

TABLE 1—Profile of Smoke-Free Restaurant and Comparison Cities

Smoke-Free City and Matched '

Comparison City

Aspen, Colo Vail, Colo

Auburn, Calif Oroville, Calif

Beverly Hills, Calif Santa Monica, Calif

Bellflower, Calif Lakewood, Calif

El Cerrito, Calif San Pablo, Calif

Lodi, Calif Merced, Calif

Martinez, Calif Pleasant Hill, Calif

Palo Alto, Calif Mountain View, Calif

Paradise, Calif Red Bluff, Calif

Roseville, Calif Chico, Calif

Ross, Calif Pt. Arena, Calif

Sacramento, Calif Fresno, Calif

San Luis Obispo, Calif Santa Maria, Calif

Snowmass, Colo Breckenridge, Colo

Telluride, Colo Steamboat Springs, Colo

1989 Population^

5 049 3 659

10 592 11 960

31 971 86 905

61 815 73 000

22 869 25 158

51 874 56 216

32 038 31 585

55 544 67 460

25 408 12 363

44 685 40 076

2 180 428

369 365 354 202

41 958 61 284

1 426 1 285

1 292 6 695

Location"

0 0

1 1

1 R

1 1

R R

R 0

Type of Smoking

Restriction'^

100% Some

100% None

100% Some

100% None

100% Some

100% None

1989 Median Household Income^

37 467 41 211

37 272 16614

54 348 35 997

32 711 44 700

39 538 25 479

30 739 24 727

45 964 46 885

55 333 42 431

22 954 19 474

39 975 19 005

84 414 21 250

28 183 24 923

25 982 29 492

39 107 33 259

31 968 29 363

=Data are from US Census Bureau.^' ""O = outside urban area; 1 = inside urban area; R = rural, nonfarm area. •̂ ''Some" refers to no more than 60% seating area for nonsmokers. "Data are from Pierce et al.22 (California) and Centers for Disease ControP^ (Colorado). ^Number of months for which data were available for this study.

%0f Smokers'*

23.5 23.5

24.1 23.6

21.8 21.8

22.9 22.9

24.1 25.1

22.0 22.0

19.7 19.7

23.6 23.6

24.1 23.6

21.6 23.6

25.2 25.1

18.9 18.9

23.5 23.5

Date Ordinance in Effect

8/85

4/91

4/87-7/87

3/91-6/92

11/91

12/90

3/92

9/92

8/91

10/91

1/90

5/92

8/90

5/89

4/88

No. Months

in Effect"

119

or not a smoke-free restaurant law was in force; and, for Colorado, a dummy variable for the winter tourist season. The regression coefficient bi quantifies the magnitude of the effect of the ordinance. In addition to analyzing data for each city separately, we pooled all data on restau- rant sales as a percentage of total retail sales for all 15 cities with ordinances for the entire year in a single analysis, including dummy variables to allow for between-city differences.

For analysis of the smoke-free bar ordinances, in addition to the analyses just described, we evaluated sales for eating and drinking establishments with liquor licenses (category 36) as a fraction of all retail sales and as a fraction of all sales by

eating and drinking establishments, using the same procedures just described, by computing

Bar Sales

Total Eating and Drinking Places Sales

In a few cases, the residuals showed evidence of a positive serial correlation (evidenced by a statistically significant Durbin-Watson statistic). The residual plots suggested a long-term nonlinear relationship that may have reflected the business cycle in California, which was strong near the beginning and end of our study period and in recession in the middle of the study period (early 1990s).

We reanalyzed the data, including a quadratic term in time (after centering the time variable to reduce the structural multicollinearity'^). P < .05 is considered statistically significant.

Results

Table 3 summarizes the results for total restaurant sales as a fraction of all retail sales and for the ratio between total restaurant sales in cities with ordinances and sales in the matched comparison cities. The first column in the table is the mean value observed over the study period.

Smoke-free ordinances generally had no statistically significant effect, either on the fraction of total retail sales that went to

1688 American Joumal of Public Health October 1997, Vol. 87, No. 10

Public Health Briefs

TABLE 2—Profile of Smoke-Free Bar and Comparison Cities and Counties, California

Smoke-Free City or County and Comparison

City or County

Anderson' Red Bluff

Davis Chico

Redding Healdsburg

San Luis Obispo (None available)

Shasta County Butte County

Santa Clara County Alameda County

Tiburon Sausalito

1989 Popuiation^

8 299 12 363

46 209 40 079

66 462 60 471

41 958

72 275 98 625

106 183 119 882

7 532 7 152

"Data are from US Census Bureau,2' •"O = outside urban area; 1 = inside urban area

Location''

0 0

1 1

R R

1 i

1 1

Type of Smoi<ing

Restriction'^

100% None

100%

100% None

: R = rurai, nonfarm area.

1989 Median Household

Income^

22 321 19 474

29 044 19 005

25 828 33 712

41 676

25 581 22 776

48 115 37 544

75 864 60 471

%0f Smokers'*

23,6

18,9

23,6

19,7

21,7

^"None" refers to locaiities in wiiich bars are specificaiiy exempted from or not mentioned in any existing ordinance.

"uata are trom rierce et ai, ^Number of months for which data were available for this study. 'County ordinance is enforced In city.

Date Ordinance

In Effect

2/93

3/93

2/93

8/90

2/93

2/94

11/93

No, Months

in Effecf=

restaurants or on the ratio between sales in smoke-free cities and sales in comparison cities (Table 3 and Figure 1). The linear model indicates that the fraction of total retail sales that went to restaurants in- creased in two cities (Bellflower and Martinez) and decreased in two cities (Paradise and Roseville). Restaurant sales relative to sales in the comparison city increased in one city (Palo Alto) and decreased in another (Paradise). These results are similar to those we previously reported.^ The nonlinear model produced similar estimates for the ordinance effects; two cities showed an increase and one a significant decrease in terms of restaurant revenues as a fraction of retail sales, and one city showed a significant increase in terms of the ratio between its sales and those of its comparison city (Table 3). Analysis of all the data in pooled regres- sions did not result in significant changes when either model was used.

The results of the analysis of the bar data appear in Table 4 and Figure 2, The linear model indicates that there were no significant effects of the smoke-free ordi- nances on bar sales as a fraction of total retail sales, on the ratio between bar sales in cities with ordinances and sales in comparison cities, or on the fraction of all eating and drinking place revenues

reported by establishments that sell all types of liquor (category 36). The nonlin- ear model indicated that there was one significant drop in sales (in Davis relative to its comparison city). Analysis of all the data in pooled regressions did not result in significant changes in any variable when either model was used.

The nonlinear model resolved the few statistically significant serial correla- tions we observed, but it had little effect on the estimates of the ordinance effects. This result suggests that the estimates of the ordinance effects are not artifacts of the model specification. The lack of consistent response suggests that the few statistically significant changes we esti- mated may simply reflect random varia- tion, given the large number of P values that were computed, rather than a system- atic effect of the ordinances.

Discussion

This study expands and confirms our earlier work showing that smoke-free restaurant ordinances do not affect restau- rant revenues. It also shows that the same is true for smoke-free bar ordinances. The cities and counties with smoke-free bar ordinances are diverse. Anderson and Redding are isolated cities within a

predominantly agricultural region of Cali- fomia. Davis is a university town. Tiburon is an affluent suburban community that enjoys heavy tourist business. San Luis Obispo is a coastal community that has a major college as well as substantial tourism. The two smoke-free counties, Shasta and Santa Clara, have ordinances that cover unincorporated areas; Shasta is rural and Santa Clara is a suburban county in the San Francisco Bay Area,

Our earlier work attracted criticism from the tobacco industry,'^ acting through Philip Morris' National Smokers Alli- ance.'^ The criticisms have included claims that there were errors in the effective dates of the ordinances, that we mischaracterized the ordinances as smoke- free when they were not, and that sales tax data are not accurate (Evans MK, A review of "The effect of ordinances requiring smoke-free restaurants on restau- rant sales" by Stanton A. Glantz and Lisa R. A. Smith. March 1997, Unpublished,). Correcting the effective dates of these ordinances does not affect the conclusions in our earlier paper.'^ A careful review of the ordinances for both the smoke-free and comparison cities shows that they meet our stated criteria in both cases.' Sales tax data include all restaurant and bar sales and are collected by an agency

October 1997, Vol. 87, No, 10 American Journal of Public Health 1689

Public Health Briefs

TABLE 3—Effect of Smoke-Free Restaurant Ordinances on Total Restaurant

City

Aspen, Colo

Auburn, Calif

Bellflower, Calif

Beverly Hills, Calif

El Cerrito, Calif

Lodi, Calif

Martinez, Calif

Palo Alto, Calif

Paradise, Calif

Roseville, Calif

Ross, Calif

Sacramento, Calif

San Luis Obispo, Calif

Snowmass, Colo

Telluride, Colo

All combined

Mean

Sales

Effect of Ordinance

Change, bi.

Fraction of total retail saies,

24.9

8.6

12.6

13.0

12.8

11.3

11.0

16.2

14.6

6.8

48.9

14.1

13.0

45.3

30.1

18.9

- 0 . 3 - 0 . 4 - 0 . 1 - 0 . 1

2.4 2.0 0.8 0.3 0.0 0.3 0.0 0.0 2.5 4.4 0.0 0.2

- 2 . 0 - 2 . 2 - 1 . 1 : - 1 . 0 :

- 1 0 . 9 : - 3 . 7 :

1.1 : 1.2: 0 . 0 : 0.1 :

- 4 . 2 : - 3 . 9 :

2 . 8 : 13.5: 0 . 0 : 0 . 0 :

± 1.1 ± 1.7 ±0.8 ±0.5 ±0.4 ±0.4 ± 1.4 ± 1.3 ±0.7 ±0.7 ±0.5 ± 0.5 ± 1.0 ± 3.2 ± 1.0 ±0.1 ±0.8 ±0.8 tO.4 ±0.4 ±5.9 ±7.2 ±0.6 t 0.6 tO.5 tO.5 t 4 . 3 t 5 . 8 t 3.7 t 5.0 tO.6 tO.6

.820

.809

.865

.805

.000

.564

.794

.964

.703

.954

.896

.021

.182

.988

.896

.015

.009

.006

.012

.070

.609

.066

.076

.953

.789

.337

.506

.488

.011

.964

.879

Model

.650

.640

.834

.503

.548

.043

.192

.052

.095

.115

.153

.267

.705

.176

.177

.264

.288

.402

.423

.543

.580

.145

.147

.256

.267

.530

.062

.230

.890

.891

.000

.000^

.000

.428

.042

.350

.278

.092

.095

.007

.000

.023

.058

.003

.006

.000

.048

.106

.003

.008

.000

.482

.042

.000

Ratio between sales in smoke-free city and sales in comparison city

Aspen, Colo

Auburn, Calif

Bellflower, Calif

Beverly Hills, Calif

El Cerrito, Calif

Lodi, Calif

Martinez, Calif

Palo Alto, Calif

Paradise, Calif

Roseville, Calif

Ross, Calif

Sacramento, Calif

San Luis Obispo, Calif

Snowmass, Colo

Telluride, Colo

All combined

1.17

.44

.50

.56

1.32

.86

.44

1.71

.69

.68

.60

1.10

1.12

.80

.43

.85

. 0 9 :

. 3 2 :

. 0 3 :

t .12 t .18 t .02

.04 ± .02 - . 0 1 j

.01 :i - . 0 6 d - . 0 6 d

.08 d - . 0 5 d - . 0 1 d

.00 d

.04 d

.27 d

.07 d - . 0 9 d - . 0 7 d - . 0 2 d - . 0 6 d -.12 i 1.93 d -.03 d -.05 d -.08 d -.08 d -.13 :• -.27 1

.04 1

.09 1 -.01 1 -.01 1

t .02 t .02 t .04 t .04 :.O8 : .10 :.O3 :.O3 : .03 : .03 : .07 : .09 :.O4 : .04 : .03 : .03 : .42 : .62 : .03 : .04 :.O5 :.O5 : .14 : .18 : .05 : .08 ::03

.03

Note. The first row for each city shows results of linear time results of quadratic time model.

^Significant positive serial correlation of residuals.

.459

.083

.170

.844

.726

.698

.158

.178

.333

.606

.834

.921

.120

.182

.001

.421

.018

.119

.511

.081

.769

.008

.322

.211

.142

.154

.342

.138

.456

.266

.627

.589

model; the

.223

.262

.077

.370

.028

.029

.279

.283

.089

.221

.554

.572

.705

.349

.479

.329

.355

.042

.173

.564

.784

.310

.324

.158

.179

.699

.710

.283

.297

.802

.805

.003

.002

.267^

.002

.629

.814

.005

.015

.214

.148

.000

.001

.000

.001«

.003

.490

.104

.002

.000

.002

.005

.059

.094

.000

.005

.010

.000

second row shows

with no interest in the effects of smoke- free ordinances. (For a detailed response to the Evans critique, contact the authors. This response is also posted on the World Wide Web at http://www.tobacco.org/ Misc/evansresponse.html.)

As noted above, an exemption for the bar area of a restaurant did not disqualify a smoke-free restaurant ordi- nance from our study of smoke-free restaurant ordinances,^ so long as the eating areas were smoke-free. The present study shows that smoke-free bar ordi- nances do not affect bar revenues. It is important to note that our analysis of the smoke-free bar ordinances relied on data for establishments with full liquor licenses (category 36). This category includes free-standing bars and bars within restau- rants. It is not possible to analyze the effects of these ordinances on these two subcategories of business separately. Nev- ertheless, it is important to evaluate the effects of smoke-free ordinances on this category of eating and drinking establish- ments because of claims by the tobacco industry that smoke-free ordinances par- ticularly affect sales of establishments that sell liquor. Moreover, the fact that some of the restaurant ordinances permitted smok- ing in bar areas does not support the tobacco industry's assertion that the lack of change in restaurant revenues in the cities shown in Table 1 is due to a shift of business to bars and the bar areas of restaurants.

The fact that we did not observe changes in the fraction of eating and drinking establishment revenues going to category 36 businesses is evidence that these ordinances do not cause shifts between types of business. It is also important to emphasize that the purpose of this study, like that of our earlier study, was to address the claim that smoke-free ordinances substantially decrease rev- enues across the board; the usual claim is a reduction of 30%. The data do not support this claim. (Our analytic methods have a power exceeding .99 to detect a 30% change in restaurant or bar revenues with a = .05.)

Food service workers enjoy less protection from secondhand tobacco smoke than any other group of employ- e e s . " Legislators and govemment offi- cials can enact health and safety regula- tions to protect patrons and employees^" in restaurants and bars from the toxins in secondhand tobacco smoke without fear of adverse economic consequences. D

1690 American Joumal of Public Health October 1997, Vol. 87, No. 10

Aubum - ^

I 9 W ia«6 1990 1992 1994 1996

w,i« a

BeDflower

1916 1966 1990 1992 19M 1996

15 <

5 Beverty Hiils

19S6 1961 1990 1992 1994 1996

a.-a. >W ^ a.af^^»^^f**—^^^f*

El Cerrito

1966 1966 1990 1992 1994 1996

5 Lodi

1966 1966 1990 1992 1994 1996

15-

S Martinez

1966 1966

15-

5 Palo Alto

1966 1966

5 Paradise

1966 1966

Roseville

teJVUeaeee

1966 1966

1990

pnn ft

1990

1992

1994

1996

15 (

>%,

1966

40 (

1962

6 0 (

1966

Sacramento

1966 1990

San Luis Obispo

1966 1990

Aspen

1964 1966 1966

Snowmass

1966 1990

Telluride

1966 1990

t^ote. The quarters in which 100% smoke-free restaurant ordinances were in effect are shown as solid points.

FiGURE 1—Restaurant saies as a percentage of total ordinances inciuded in this study.

retaii

1992

1990

fut

1992

• tLJt

1992

saies for the 15 communities with smoke-free

1994 1996

1992 1994 1996

1994 1996

restaurant

Anderson Redding Shasta County

1992

Davis

1990

10 T

1966 1969 1990 1992 1994 19

San Luis Obispo

Tiburon

1992 1966 1966 1990 1992 1994 1996

Santa Clara County

1966 1966 1990 1992 1994 1996

Note. The quarters in which 100% smoke-free bar ordinances were in effect are shown as solid points.

FIGURE 2—Bar sales as a percentage of total retail sales for the 7 California communities with smoi<e-free bar ordinances inciuded in this study.

October 1997, Vol. 87, No. 10 - American Journal of Public Health 1691

Public Health Briefs

TABLE 4—Effects of Smoke-Free Bar Ordinances on Total Bar Sales, California

City

Anderson

Davis

Redding

San Luis Obispo

Santa Ciara County

Shasta County

Tiburon

Ail combined

i\/1ean

Effect of Ordinance

Change, bi. P

Fraction of total retail sales, %

3.1

3.3

2.6

3.8

4.5

2.4

41.3

7.1

- 0 . 7 ± - 0 . 4 ±

0.1 ± 0.0 ± 0.1 ± 0.0 ± 0.0 ± 0.0 ± 0.0 ± 0.0 ± 0.2 ± 0.2 ± 0.6 ± 0.3 ± 0.5 ± 0.3 ±

0.8 0.5 0.8 0.2 0.2 0.2 0.3 0.3 0.2 0.2 0.1 0.1 5.1 0.6 0.4 0.5

.404

.514

.752

.943

.456

.731

.907

.945

.775

.996

.076

.270

.904

.964

.258

.456

.043

.613

.658

.718

.098

.129

.052

.053

.135

.137

.505

.508

.103

.105

.979

Modei

.689=

.001

.000

.417

.516

.374

.574

.059

.129

.000

.396

.610

.000

Ratio between sales in smoke-free city and sales in comparison city

Anderson

Davis

Redding

San Luis Obispo

Santa Clara County

Shasta County

Tiburon

Ail combined

Anderson

Davis

Redding

San Luis Obispo

Santa Clara County

Shasta County

Tiburon

Ail combined

.49

.42

18.05

.31

1.52

1.07

0.70

2.76

- . 1 5 ± - . 2 7 ± - . 0 8 ± - . 1 0 +

.66 + -7.41 ±

.05 ±

.04 ±

.02 ± - . 0 2 ± - . 1 0 ± - . 0 9 ± - . 2 4 ±

.21 ±

.12

.11

.04

.04 5.3 3.91

.06

.09

.01

.08

.06

.07

.45

.53

.224

.033

.073

.021

.902

.076

.411

.664

.067

.757

.134

.205

.589

.695

Fraction of total restaurant sales, %

19.9

17.9

26.6

29.3

34.8

23.9

76.6

.32

4.8 ± - 3 . 2 ± - 1 . 5 ± - 1 . 6 ± - 0 . 9 ± - 1 . 1 ±

0.0 ± 1.0 ± 0.9 + 0.1 ± 1.5 + 0.5 ±

- 1 . 6 ± - 1 . 7 ±

0.6 ± 0.4 ±

5.5 3.4 1.1 1.1 0.9 1.1 0.3 2.6 0.5 0.7 0.8 1.0 3.1 3.4 0.8 1.0

.401

.358

.194

.147

.373

.331

.907

.704

.065

.890

.067

.597

.603

.623

.505

.692

.153

.400

.696

.767

.023

.596

.486

.576

.601

.411

.413

.898

.903

.042

.711

.739

.753

.533

.539

.052

.197

.809

.827

.576

.603

.036

.955

.244=

.038

.000

.823=

.002

.000

.011

.032

.000

.692

.000

.002

.005

.374

.046

.000

.000*

.000

.732

.894

.000

Note. For each city, the first row of data shows resuits of the iinear time modei; the second row shows results of the quadratic time modei.

^Significant positive serial correiation of residuals.

Acknowledgment This work was supported by National Cancer Institute grant CA-61021.

References 1. Local ordinance database. Berkeley, Calif:

Americans for Nonsmokers Rights; 1997.

2. Macdonald H, Glantz S. The political realities of statewide smoking legislation: the passage of California's Assembly Bill 13. Tobacco Control. 1997;6:41-54.

3. Glantz S, Smith LRA. The effect of ordinances requiring smoke-free restau- rants on restaurant sales. Am J Public Health. 1994;84:1O81-IO85.

4. The San Luis Obispo Smoking Ordinance: A Study of the Economic Impacts on San Luis Obispo Restaurants and Bars. San Luis Obispo, Calif: Taylor Consulting Group (under contract to the City of San Luis Obispo); January 1993.

5. Maroney N, Sherwood D, Stubblebin W. The Impact of Tobacco Control Ordinances on Restaurant Revenues in California. Claremont, Calif: The Claremont Institute for Economic Policy Studies and the Claremont Graduate School; January 1994.

6. Centers for Disease Control. Assessment of the impact of a 100% smoke-free ordi- nance on restaurant sales—West Lake Hill, Texas, 1992-1994. MMWR Morb Mortal Wkly Rep. 1995;44:370-372.

7. Pope G, Bartosch W. Effect of local smoke-free restaurant policies on restau- rant revenue in Massachusetts. Waltham, Mass: Center for Health Economics Re- search (conducted for Massachusetts De- partment of Public Health); April 22, 1997.

8. Sciacca J, Ratliff M. Prohibiting smoking in restaurants: effects on restaurant sales. Am J Health Promotion. In press.

9. Samuels B, Glantz S. The politics of local tobacco control. JAMA. 199I;266:211O- 2117.

10. Samuels B, Begay M, Hazan A, Glantz S. Philip Morris' failed experiment in Pitts- burgh. J Health Polit Policy Law. 1992; 17: 329-351.

11. Traynor MP, Begay ME, Glantz SA. New tobacco industry strategy to prevent local tobacco control. JAMA. 1993;270:479- 486.

12. Taxable Sales in California (Sates and Use Tax) (Table 5). Sacramento, Calif: State of Califomia Board of Equalization; 1986- 1996.

13. Sales Tax Statistics for Counties and Selected Cities and Detailed Industry Reports by County. Denver, Colo: State of Colorado Department of Revenue; January 1982-June 1996.

14. Glantz S, Smith L. The Effect of Ordi- nances Requiring Smoke-Free Restaurants on Restaurant Sates in California. San Francisco, Calif: University of Califomia, San Francisco, Institute for Health Policy Studies; 1992. Monograph series.

15. Glantz SA, Slinker BK. Primer of Applied Regression and Analysis of Variance. New York, NY: McGraw-Hill; 1990.

16. Richards B. Pro-tobacco groups step up attacks on a longtime foe. Wall Street J. July 23, 1997:B1.

17. Stauber J, Rampton S. Toxic Sludge Is Good for You. Monroe, Me: Common Courage Press; 1995:14,29-31.

18. Erratum. In: Glantz S, Smith L. The effect of ordinances requiring smoke-free restau- rants on restaurant sales. Am J Public Health. 1997;87:1729-1730.

19. Gerlach K, Shopland D, Hartman A, Gibson J, Pechacek T. Workplace smoking policies in the United States: results from a

1692 American Joumal of Public Health October 1997, Vol. 87, No. 10

Public Health Briefs

nationrfl survey 'of over 100,000 workers. Tobacco Control. In press.

20. Siegei M. Involuntary smoking in the restau- rant workplace: a review of employee expo- sure and health effects. JAMA. 1993;270:490- 493.

21. United States Census of Population and Housing 1990: Summary Tape File 3A. Washington, DC: US Dept of Commerce, Bureau ofthe Census; 1990.

22. Pierce J, Frakas A, Evans N. Tobacco Use in Califomia 1992: A Focus on Preventing

Uptake in Adolescents. Sacramento, Calif: Califomia Department of Health Services; 1993.

23. Behavioral Risk Factor Surveillance Sys- tem—Summary Prevalence Report. Atlanta, Ga: Centers for Disease Control; 1991.

Thyroxine Values from Newborn Screening of 919 Infants Bom before 29 Weeks' Gestation

A B S T R A C T

Objectives. Severe transient hy- pothyroxinemia in premature infants is associated with cerebral palsy and mental retardation; this study as- sessed its prevalence in very prema- ture infants.

Methods. Congenital hypothy- roidism screening programs in three states provided thyroxine values for 919 newborn infants younger than 29 weeks who were enrolled in a multicenter study.

Results. Thyroxine values were lower than 4.0 |ig/dL in 2 1 % of survivors and increased each week by 0.6 |ag/dL (95% confidence inter- val [CI] = 0.4, 0.7). At tests done 1 to 2 days after birth, levels were 2.5 |ig/dL higher (95% CI = 1.8, 3.3) than at tests done at 8 to 14 days. In New York, levels were 1.0 (ig/dL higher (95% CI = 0.3,1.6) than else- where. The levels of infants who died were 1.3 ng/dL lower (95% CI = 0.6, 2.0) than those of survivors.

Conclusions. Severe transient hypothyroxinemia is common in very premature infants and deserves fur- ther study. (Am J Public Health. 1997;87:1693-1697)

M. Lynne Reuss, MD, MPH, Alan Leviton, MD, SM, Nigel Paneth, MD, MPH, and Mervyn Susser, MB, BCh, FRCP(E), DrPH

Introduction Preterm infants often have low thy-

roxine levels postnatally, a condition referred to as transient hypothyroxinemia of prematurity.'"'^ Transient hypothyrox- inemia of prematurity is a self-limited phenomenon thought to be caused by immaturity of the hypothalamic-pituitary- thyroid system and by changes in thyroid function that accompany severe illness, that is, nonthyroidal illness. Congenital hypothyroidism is not thought to explain why transient hypothyroxinemia is de- tected at newbom screening of premature infants because thyrotropin levels are normal. However, recent studies of pre- term infants have linked very low thyrox- ine levels with abnormal cognitive and neurological development at ages 2 through 9 years.'"*"'* It has been difficult to establish what represents a very low thyroxine level at any given gestational age because little is known about the gestational age-specific distribution of thyroxine values in very preterm infants. State screening programs tend to collect and report infonnation classified by birth- weight, not by gestational age,'^ and they rarely report quantitative results.

In this paper we describe thyroxine- screening findings in 919 preterm infants bom before 29 weeks' gestation and enrolled in a multicenter study of cranial ultrasonographic abnormalities, the Devel- opmental Epidemiology Network Study. These infants, whose gestational ages were established according to a study protocol, received intensive neonatal care in one of four nurseries in three states: Massachusetts, New Jersey, and New York. Quantitative thyroxine-screening results were obtained from state congeni-

tal hypothyroidism-screening programs and were assessed in relation to survival, postmenstrual and postnatal age at the time of screening, and site of care.

Methods

From January 1991 through Decem- ber 1993, 1662 infants weighing 500 through 1500 g were systematically en- rolled in a multicenter study of neonatal brain injury, the Developmental Epidemi- ology Network Study. Study infants were bom in four hospitals in Massachusetts, New Jersey, and New York (two hospi- tals). Of the 919 bom at less than 29 weeks' gestation, and therefore at high risk for severe transient hypothyrox- inemia of prematurity, 746 survived to discharge from the intensive care nursery.

Gestational-age estimates were based on fetal ultrasound obtained before the 14th week of gestation (32%), dates in the prenatal record (62%), matemal postpartum interview (4%), and the admission logbook ofthe neonatal intensive care unit (2%).

At the time of the study, M. Lynne Reuss was with the Sergievsky Center, Columbia Univer- sity, New York, NY. She is now with the Believue Research Foundation, Niskayuna, NY. Alan Leviton is with Harvard Medical School, Boston, Mass. Nigel Paneth Is with the College of Human Medicine, Michigan State University, East Lan- sing. Mervyn Susser is with Columbia Univer- sity, New York, NY.

Requests for reprints should be sent to M. Lynne Reuss, MD, MPH, Believue Research Foundation, 2210 Schenectady Troy Rd, Niska- yuna, NY 12309.

This paper was accepted November 8, 1996.

Editor's Note. Dr Heinz Berendes served as the responsible editor and Dr Mary Northridge as editor for this paper. As is our practice, Dr Mervyn Susser had no part in the review and decision process.

Octoberl997,Vol. 87, No. 10 American Journal of Public Health 1693